Feeling travel photographs: Emotional labelling in machine-generated semantic differentials

Feeling travel photographs: Emotional labelling in machine-generated semantic differentials | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Feeling travel photographs: Emotional labelling in machine-generated semantic differentials Joanne Yu, Denis Helic, Astrid Dickinger, Markus Strohmaier This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7591283/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 29 Apr, 2026 Read the published version in Information Technology & Tourism → Version 1 posted 9 You are reading this latest preprint version Abstract The pervasive rise of social media has placed visual content, especially photographs, at the forefront of digital marketing in tourism. As marketers strategically leverage emotionally resonant content on platforms, this research explores the relationship between hedonic and utilitarian dimensions embedded in pictorial representations and their impact on user engagement in tourism. By utilising a machine-generated semantic differential, this study analyses 33,001 picture-based posts from popular destinations on Instagram. The findings of this study reveal that user engagement dynamics vary across themes, with multimodal excursions (e.g., modes of transportation) emphasising informativeness over fun and thrill, while art and sculpture benefit from the integration of functionality. By using an interdisciplinary approach, this research explores the emotional connotations of visual content on the engagement of potential tourists. Overall, this paper underscores the importance of tailoring content strategies to specific thematic contexts and challenges preconceived notions about the universal appeal of certain pictorial elements. semantic differential visual analysis hedonic consumption utilitarian consumption machine learning Figures Figure 1 Figure 2 Figure 3 Figure 4 1. Introduction The rise of social media has led to a growing emphasis on visualisation (Xu et al., 2023 ). Particularly, photographs have become the backbone, with platforms such as Instagram relying heavily on visual content to attract and engage with users (Hauser et al., 2022 ). Driven by the evolution of experience design, marketers now focus on creating visually engaging content that elicits emotional arousal in a hedonic context (Gon, 2020 ). Among various types of visually centred multimedia, such as images, livestreaming, and videos (Fu et al., 2024 ), pictures play a crucial role (Yu & Egger, 2021 ). In the tourism context, pictures are key to understanding the emotional experiences conveyed through visual elements (Vu et al., 2025 ). Visual content not only showcases product features and service details, but also effectively conveys hedonic values that elicit emotional responses (Deng & Li, 2018 ). For instance, visual elements of tourism pictures such as colours and objects have specific associations that can evoke diverse feelings in viewers (López-Chao & Lopez-Pena, 2020 ; Yu & Egger, 2021 ). In fact, the discussion above aligns with the broader dichotomy of hedonic and utilitarian experiences in consumption emotions (Davari et al., 2024 ). In tourism, hedonic consumption refers to experiences that provide pleasure and emotional satisfaction (Soldat et al., 2024 ), such as the enjoyment of beautiful scenery or cultural activities. Utilitarian consumption, on the other hand, focuses on practicality and functional benefits (Lee et al., 2021 ), like the convenience and efficiency of travel services. Balancing practical tourism information (Hong et al., 2023 ) with the pleasure-seeking aspect of social media and tourism (Virtanen et al., 2017 ) is salient. This balance affects how tourists engage with visual content and interact with digital media (Li & Xie, 2020 ). Nonetheless, despite the increasing interest in understanding how photographs contribute to overall experiences (Picazo & Moreno-Gil, 2019 ), most research in this domain centres on the hedonic aspects within the realm of tourism marketing (Filieri et al., 2021 ; Yu & Egger, 2021 ). Although the experiential impact of visuals in tourism marketing has been recognised (Yu & Egger, 2021 ), the functional angles, often encapsulated in utilitarian value, remain underexplored. Furthermore, recent scholars have underpinned that emotional aspects within the tourism imagery are still not thoroughly explored, mainly due to the challenge of objectively measuring the intangible nature of emotions (Vu et al., 2025 ). Methodologically speaking, previous studies on hedonic and utilitarian values primarily relied on semantic differential techniques with traditional survey methods (Kuo et al., 2022 ; Lee et al., 2021 ; Li et al., 2020 ). While useful for capturing opinions on bipolar scales, these surveys face criticisms like social desirability bias and central tendency errors (Brühlmann et al., 2020 ). More advanced approaches involve data analytics to examine how visual elements influence user engagement on social media (Rietveld et al., 2020 ). Despite advancements in visual data analytics, researchers often rely on previous literature for interpretation (Yu & Egger, 2021 ) and focus mainly on basic emotional responses (Vu et al., 2025 ). This highlights the need for a new approach to better understand the experiences conveyed by visual elements (Mathew et al., 2020 ). One way to measure the underlying emotions in visual content is the use of machine learning to generate interpretable bipolar dimensions (Reelfs et al., 2022 ). This cutting-edge technique in data analytics (Engler et al., 2023 ) shows promise for understanding experiential marketing of tourism destinations and informing effective strategies. Therefore, this research aims to investigate the intertwining of hedonic and utilitarian aspects of tourism-related pictures shared by marketers and their relationship with user engagement using machine-generated semantic differentials. By bridging data analytics and visual content marketing, this research is novel in that it uses an interdisciplinary approach to examine the emotional connotations of pictures on user engagement in social media. Specifically, this study advances the literature by exploring the complex relationship between hedonic and utilitarian consumption in visual tourism marketing across various themes. Using machine-generated semantic differentials provides a new and more objective way to interpret emotional responses to images. Practically, these insights help marketers balance emotional appeal and craft effective visual strategies to boost engagement. 2. Literature review 2.1. The effect of visual content on user engagement In the broader context of multimedia marketing, text, videos, and images all play significant roles in engaging consumers (Wang & Luo, 2025 ). However, pictures have emerged as particularly influential in visual marketing due to their immediacy and emotional impact (Yu & Egger, 2021 ). The use of various pictorial elements play a critical role in the creation of visual content, as they influence how consumers perceive the images (Xu et al., 2023 ). For instance, the colour scheme (Yu & Egger, 2021 ), message appeals (Rietveld et al., 2020 ), and image quality (Li & Xie, 2020 ) can significantly impact the audience’s emotional response and engagement in digital communication. Building on the principles of experience design, marketers are now striving to intentionally create content and incorporate design thinking (Tussyadiah, 2014 ) when developing marketing materials. Through carefully crafted visual content, marketers seek to create a more meaningful and authentic relationship with online users (Wies et al., 2023 ), thereby fostering emotional connections that can lead to increased engagement (Lim et al., 2020 ). Particularly, Instagram has become an essential platform for tourism marketing (Hauser et al., 2022 ). The platform is known for its heavy reliance on visuals, with photos and videos being the primary means of communication. On Instagram, evaluation of marketing campaigns relies on user engagement, which reflects how well a post resonates with potential tourists (Yu & Egger, 2021 ). The observable metrics is typically shown by the number of likes and comments (Rietveld et al., 2020 ). Notably, differentiating likes and comments as dimensions of user engagement is crucial because they infer various level of interaction and interest that users have (Wies et al., 2023 ). Interestingly, recent research revealed that pictures featuring a special or unique activity tend to receive more likes, whereas content requiring a higher level of awareness (e.g., environmental issues) generates more comments (Aramendia-Muneta et al., 2020 ). In the context of sport events, embedding cultural elements in the pictures lead to higher engagement, while nature-related pictures yield the fewest likes and comments (Taberner & Juncà, 2021 ). However, these studies primarily focus on the surface-level objects within images (Taecharungroj & Mathayomchan, 2021 ; Yu et al., 2025 ), leaving the emotional dimensions embedded in these image objects across different contexts underexplored. Scholars have suggested that visual content can indeed convey a wide range of emotional experiences (Yu & Egger, 2021 ). For instance, photos of serene beach scenes may evoke peace and relaxation, while images of waves can generate excitement or even fear (Vu et al., 2025 ). Seeing that digital experiences increasingly focus on eliciting emotional responses, it is no longer sufficient for visual elements to be merely appealing (Hauser et al., 2022 ). Instead, they need to (implicitly) convey emotional experiences that enable potential tourists to resonate and form a connection with the content (Xiao et al., 2020 ). 2.2. Hedonic and utilitarian consumption in digital marketing The evaluation of emotional dimensions in visual content aligns seamlessly with the framework of consumption values, encompassing hedonic and utilitarian experiences (Davari et al., 2024 ). Hedonic experiences are experiential and enjoyable in nature (Soldat et al., 2024 ), such as the feelings of excitement or relaxation that consumers derive from leisure activities or holiday (Su et al., 2024 ). In contrast, utilitarian consumption centres around practical benefits and functionality, emphasising aspects such as usefulness and efficiency (Roggeveen et al., 2015 ). Utilitarian values also involve an appreciation for the practical benefits by assessing whether a product or service can aid in achieving specific goals (Lee et al., 2021 ). Revisiting foundational consumption theories highlights the critical role they play in evaluating how visuals offer marketers opportunities to showcase their products and services. This can be achieved through diverse presentation forms that leverage both hedonic and utilitarian approaches (Su et al., 2024 ). While it appears that there has been a shift towards a more experiential and emotional emphasis in tourism, as consumers increasingly seek out hedonic value and deeper emotional attachments to enhance the overall experience (Tussyadiah, 2014 ), it is important to acknowledge that many experiences concurrently serve functional purposes. Taking transportation as an illustrative example, transportation services, such as those offered by ride-sharing platforms or travel agencies, often highlight not only the joy of the journey but the utilitarian efficiency and convenience (Yuda Bakti et al., 2020 ). On social media, even as companies tap into the hedonic aspect of experience (e.g., happiness, excitement) to craft successful marketing campaigns (Liadeli et al., 2023 ), functional elements in digital materials are not to be underestimated (Chen & Fu, 2018 ). Certainly, digital content mainly serving functional purposes may be less effective in triggering engagement (Liadeli et al., 2023 ). Yet, although visuals might emphasise the hedonic allure of an environment, it simultaneously addresses practical considerations (Chen & Fu, 2018 ) such as comfort, amenities, and accessibility. Hence, the evaluation of these emotional nuances becomes integral to comprehending how consumers extract value from their interactions with diverse forms of visual content. To do so, the emergence of visual analytics such as image mining (Ma & Palacios, 2021 ), clustering (Arefieva et al., 2021 ), and segmentation (Nguyen et al., 2018 ), has become increasingly important by providing an effective way that enables marketers to interpret visual data (Aramendia-Muneta et al., 2020 ). 2.3. Machine-generated semantic differentials Rooted in psychology, one way to interpret the data is the application of the semantic differential (Mindak, 1961 ), which is used to measure and analyse one’s perceptions, attitudes, and emotions towards various stimuli such as products, services, and brands (Huang et al., 2012 ). Conventionally, this technique involves asking individuals to rate the concepts being tested along a series of bipolar scales (Mathew et al., 2020 ), such as good-bad and strong-weak, which offers insights into the underlying attitudes and emotions that people associate with the stimuli (Huang et al., 2012 ). One of the key advantages of the semantic differential is that it provides a structured approach to measuring subjective attitudes and perceptions (Lin et al., 2024 ). The semantic differential has been adopted extensively in marketing research to measure attitudes and preferences (Huang et al., 2021 ; Lim & Childs, 2020 ). For example, researchers have used it to analyse consumers’ perceptions of product design (Huang et al., 2012 ), to understand their attitudes towards brands (Lim & Childs, 2020 ), and to measure the effectiveness of social media advertisements (Sung et al., 2020 ). One area where the semantic differential has been particularly useful is in understanding emotions (Das & Varshneya, 2017 ; Huang et al., 2012 ) or hedonic and utilitarian value (Voss et al., 2003 ). By using bipolar scales that capture affective terms such as dull-exciting, enjoyable-unenjoyable, it is possible to obtain a better understanding of the emotional responses a person may be undergoing (Das & Varshneya, 2017 ; Voss et al., 2003 ). Given the swift accumulation of data, recent advancement in data analytics has shown promising results in interpreting pre-trained word embeddings (Mathew et al., 2020 ). Echoing semantic differential, the learning model relies on the concept of polar opposites, where words are paired with their antonyms to enable interpretability of the embeddings (Reelfs et al., 2022 ). Its application can be seen in research focusing on sentiment analysis (Engler et al., 2023 ). For example, Reelfs et al. ( 2022 ) transformed arbitrary word embeddings into interpretable counterparts for emotion detection based on visual symbols (e.g., emojis). However, one limitation is that the embedding information in the digital landscape (Mathew et al., 2020 ) may not fully capture the complexity of consumer behaviour in specific contexts. Moreover, the use of pre-trained word embeddings may not accurately reflect the unique nuances in meaning that are specific to different industries, products, and services (Mathew et al., 2020 ). Therefore, it is essential to conduct research that incorporates how different marketing context may influence consumer behaviour (Stremersch et al., 2023 ). 3. Methodology This study employs a systematic approach to explore the emotional and experiential dimensions of tourism-related imagery on social media and their impact on user engagement. The research is structured into five main steps: 1) destination selection, 2) data extraction, 3) image captioning and topic modelling, 4) machine-generated semantic differentials, and 5) analysis of user engagement. The detailed procedures follow. An overview of the methodological procedures is presented in Fig. 1 . 3.1. Step 1: Destination selection This study gathered data from the most popular destinations in different European countries. Specifically, destinations were selected based on the top 15 cities having the highest number of total bed-nights reported in the 18th edition of the CityDNA Benchmarking Report in Europe (City Destinations Alliance, 2022 ). Yet, to maximise the diversity of destination characteristics, the second-ranked city in the same country was excluded. Cities without an official Instagram account from the destination marketing organisation were also excluded from further analysis. This led to a total of ten destinations in ten different countries, as summarised in Table 1 . Table 1 Summary of the selected destinations ( as of March 2023 ) City Bednights Account Post (n) Follower (n) London 25,542,000 @visitlondon 3,981 1.6M Paris 20,468,000 @parisjetaime 3,490 687K Berlin 13,983,000 @visitberlin 2,419 444K Madrid 10,932,000 @visita_madrid 1,902 379K Stockholm 9,233,000 @visitstockholm 4,680 345K Amsterdam 5,776,000 @iamsterdam 4,464 322K Vienna 5,407,000 @vienna 3,588 501K Prague 5,257,000 @cityofprague 603 115K Milan 5,256,000 @visit_milano 2,560 168K Lisbon 5,186,000 @visit_lisboa 4,691 133K 3.2. Step 2: Data extraction Data extraction was conducted in March 2023 using Apify, a web scaping platform, for all available posts of each account on Instagram, leading to a total of 35,459 posts. The collected data included post captions, date of the post, type of post (pictures/videos), number of likes and comments, post URLs, and image URLs. After removing 2,346 video-based posts, all pictures were downloaded based on the image URLs. Yet, since some of the pictures were duplicated and around 100 images were unavailable for download, potentially because they were deleted by the respective users, those records were excluded. The final dataset consisted of 33,001 picture-based posts. 3.3. Step 3: Image captioning and topic modelling Based on the extracted pictures, the next step involves classifying visual content based on their characteristics to gain deeper insights into different tourism experiences. First, image captioning, a technique for object recognition, was conducted using Open Clip Torch (Radford et al., 2021 ), an open-source method from OpenAI. Unlike traditional image annotations that provide simple labels, this method generates short, descriptive captions for each image, such as “ two people cross a street in front of a building ”, offering a connected narrative rather than a list of disconnected labels like “ people ,” “ building ,” and “ street .” This approach enables an automatic understanding of scene context, location, objects, people, and their interactions (Hossain, 2019). Moreover, the syntactically coherent textual descriptions produced with this tool integrate seamlessly with natural language processing techniques (You, 2016) that rely on context-sensitive embeddings for subsequent analysis. Thereafter, BERTopics (Grootendorst, 2020 ) were used to extract themes from the automatically generated image captions. This process employed sentence transformers (Reimers & Gurevych, 2019 ), to create context-sensitive embeddings using a pretrained large language model (LLM) (i.e., the “all-MiniLM-L6-v2” model) due to its strong performance in natural language processing tasks like sentiment analysis, summarisation, named entity recognition, and translation (Liu, 2020). Applying these transformers to the image captions embedded each image into a 384-dimensional vector space. In this space, images with similar captions tend to be located close to each other, indicated by small angles between their vectors. These caption vectors were then utilised for topic modelling to group images into thematically similar categories and to relate them to the hedonic and utilitarian aspects of visual content (see Section 3.4). Subsequently, Uniform Manifold Approximation and Projection (UMAP) were applied to reduce the dimensionality of sentence embeddings (McInnes et al., 2018 ). These lower-dimensional representations were then used for clustering the images. Only after the clustering process was completed did the pre-processing to refine topic representations occur. This involved removing stopwords, followed by applying the k-means algorithm to group similar vectors into clusters representing topics. Next, the term frequency-inverse document frequency (TF-IDF) scores, used to evaluate word relevance, were downscaled using BERTopics, which involves taking the square root of the scores to lessen the impact of common words. Each word within a topic was assigned a class-based TF-IDF score, with the top words being those with the highest scores. Ultimately, 15 distinct topics were extracted, showing clear separation between clusters. To assess the clustering quality, the silhouette coefficient was calculated, yielding a value of 0.36, indicating a moderate level of clustering effectiveness. 3.4. Step 4: Image captioning in machine-generated semantic differentials Afterwards, this study applied semantic differential as the measurement scale, and the bipolar dimensions were visualised in a framework for each of the identified topics. The bipolar measurement is based on the hedonic/utilitarian consumption scale (Voss et al., 2003 ). The hedonic elements include fun–not fun, exciting–dull, delightful–not delightful, thrilling–not thrilling, enjoyable–unenjoyable, happy–not happy, pleasant–unpleasant, playful–not playful, cheerful–not cheerful, amusing–not amusing, sensuous–not sensuous, and funny–not funny. The utilitarian elements are effective–ineffective, helpful–unhelpful, functional–not functional, necessary/unnecessary, practical–impractical, beneficial–harmful, useful–useless, sensible–not sensible, efficient–inefficient, unproductive–productive, handy–not handy, and problem solving–not problem solving. For the positive poles, the researchers extracted their definitions from the WordNet (Fellbaum, 2005 ), a comprehensive lexical database for English. WordNet definitions are typically brief sentences; for example, “fun” is defined as “activities that are enjoyable or amusing,” and “cheerful” as “being full of or promoting cheer; having or showing good spirits.” These definitions were transformed using the same sentence transformers applied to the automatically extracted picture captions, embedding the semantic differentials as dense vector representations within a shared 384-dimensional space. Each positive pole is embedded in this space, with negative poles represented as the negatives of their positive counterparts. For each scale, the pole vectors span a subspace within this 384-dimensional space, allowing the projection of image caption vectors onto these subspaces. Each scale comprising twelve poles, resulting in a 24-dimensional subspace. By projecting the picture captions onto these semantic differential subspaces, the study generated dense vector representations that link image captions with the hedonic and utilitarian scales, embedding the captions into the same vector space defined by these scales. Notably, this study performed correlation analysis on the semantic differentials before the projection was calculated. The correlation analysis revealed weak to moderate correlations with stronger correlations within the hedonic (maximal correlation was 0.78 between “pleasant” and “enjoyable” and minimal correlation was 0.25 between “playful” and “enjoyable”) than within the utilitarian subscale (maximal correlation was 0.63 between “productive” and “effective” and minimal correlation was 0.09 between “beneficial” and “problem solving”). The correlations between the subcomponents of the different subscales were substantially weaker. Although some of the correlations can be considered strong, this study eliminated these collinearities within the subsequent regression analysis. 3.5. Step 5: Analysis of user engagement across different marketing contexts After preparing the pre-trained word vectors, the analysis proceeded to assess user engagement through multiple regression analysis. Since likes and comments reflect different levels of engagement, they were examined separately (Aramendia-Muneta et al., 2020 ). The regression analysis began by eliminating collinearities in hedonic and utilitarian subscales using the variance inflation factor (VIF), removing components with a mean VIF over four (Wu & You, 2022 ). Since likes and comments are count data, a negative binomial regression model was employed. To normalise engagement metrics, likes and comments were divided by the number of followers. In the model, this normalisation is achieved by setting an offset (exposure) parameter, which adjusts for differences in follower counts. The model then automatically accounts for this by dividing the outcome variables by the number of followers. Because the engagement data were not normally distributed, a logarithmic transformation was applied. Thus, the final regression models were multiplicative, with coefficients interpreted as relative percentage differences from baseline levels. Notably, negative binomial models assume equal mean and variance, but overdispersion often occurs when variance exceeds the mean. To address this, the overdispersion parameter was adjusted to scale the data appropriately. The study determined the optimal settings by fitting models across a range of overdispersion values and selecting the one closest to one. Lastly, the goodness-of-fit of all models was evaluated using chi-squared tests on the likelihood ratio comparing each model to the baseline. While all available data has been collected, early years showed mostly zero comments and likes. To mitigate potential bias, the subsequent analysis focused on the years from 2016 to 2022. Due to seasonality in tourism, the data were split into summer and winter seasons (summer = 0, winter = 1) to account for potential variations in user preferences. The seasonality in demand also shows impact on online engagement (Villamediana et al., 2019 ). In this research, summer spans from April through September and winter is between October and March. Next, four regression models were run altogether: for both engagement types (i.e., likes and comments) and with all 15 topics as control variables and without topics. Multicollinearity analysis included six elements from the hedonic/utilitarian scales: three hedonic (fun, thrilling, sensuous) and three utilitarian (functional, useful, productive) (see Table 4 ). 4. Results The findings are structured into four distinct sections. The first presents topic modelling results from image captions to overview tourism-related picture attributes. The second uses semantic differentials to visually represent bipolar dimensions of each topic, highlighting hedonic and utilitarian aspects. The final two sections explore the interaction effects of these facets on user engagement. 4.1. Topic modelling based on image captions Based on the BERTopics, this study reveals 15 clusters featuring different pictorial contexts. Topics were named using TF-IDF keywords and representative pictures. Table 2 shows the clusters, organised into three groups based on the inter-topic distance map: urban exploration experiences (featuring urban elements like churches and Ferris wheels), scenery and leisure retreats (focused on nature, art, and atmosphere), and multimodal excursions (involving transportation modes like cycling and railways). Table 2 Summary of the BERTopics results Cluster Topic Keywords n Group 1: Urban Exploration Experiences 0 Outdoor gathering Group, people, sitting, fountain, crowd, sidewalk, standing, walking, Eiffel, women 5,833 5 Dining ambience Table, food, plate, wine, topped, store, cup, coffee, chairs, holding 2,706 7 Church and Architecture Church, cathedral, car, cars, parked, clock, trolley, steeple, road, driving 1,722 9 Ferris wheel and festivals Christmas, Ferris, wheel, tree, decorations, rain, lit, lights, day, night 1,027 12 Castle architecture Castle, decker, double, windmill, bus, driving, clock, Brandenburger, tor, windmills 791 Group 2: Scenery and Leisure Retreats 1 Art and sculpture Woman, statue, man, person, mural, painting, standing, holding, dog, dress 4,878 2 Floral impressions Flowers, sign, plants, room, door, windows, staircase, leading ceiling, archway 3,946 3 Nature and parks Trees, sky, park, rainbow, tree, background, garden, flag, pond, flags 3,347 6 Riverside charm Sun, bridge, setting, bicycle, alley, way, parked, shining, river, bicycles 1,977 14 Cityscape and tranquil scenery Skyline, windows, tulips, sunset, square, houses, town, point, high, row 286 Group 3: Multimodal Excursions 4 Boat tours Boats, boat, docked, water, canal, body, harbor, pier, floating, aerial 3,037 8 Railway exploration Train, tracks, station, travelling, tunnel, subway, track, platform, yellow, trains 1,072 10 Cycling adventures Bike, riding, night, person, scooter, aerial, man, bicycle, street, colorful 904 11 Elevated urban transit Tram, high, point, bus, vantage, going, decker, driving, double, dome 872 13 Carriage horse rides Drawn, carriage, horse, balconies, café, horses, swan, swans, skyline, carriages 604 4.2. Bipolar dimensions of pictorial attributes Building on identified topics, the study visualises their hedonic and utilitarian aspects using a novel framework for interpreting pre-trained embeddings. By averaging embeddings per topic, diagrams show how topics are positioned in polar coordinates based on hedonic and utilitarian scales (Voss et al., 2003 ). These illustrate user reactions to specific image subjects, with each dimension representing opposing semantic pairs. Responses near the edge indicate a strong correlation, while those near the centre suggest a weaker connection. Figure 2 consists of pictorial topics related to urban experiences. Interestingly, utilitarian aspects are prominent, as images of architecture and outdoor spaces evoke handiness and accessibility. When it comes to dining ambience, the images strive to convey sensory and delightful experiences. In Fig. 3 , which depicts scenery and leisure topics, sensory consumption predominates, while purely amusing consumption is absent. An intriguing observation is that content producers appear to intentionally minimise practical feelings in nature-related topics. Turning to Fig. 4 , which encapsulates tourism experiences associated with commuting and transportation, the results unsurprisingly demonstrate an emphasis on conveying practical and problem-solving consumption. Moreover, there is a notable presence of sensuous consumption, highlighting the foundational role of sensory engagement for tourists. However, a few unexpected observations have surfaced. For instance, the results suggest that pictures associated with Ferris wheels and festivals tend to be slightly less hedonic and convey more unpleasant consumption. Similarly, images featuring natural parks and flowers appear to connote unpleasant experiences. Given the prevailing strategies employed by destination marketers in presenting diverse tourism attributes, there is a need for further exploration into their effectiveness in influencing potential tourists’ engagement with the posts, which will be addressed in the subsequent analysis. 4.3. Regression models: Initial exploration on hedonic/utilitarian scale The regression models provide an overview of the contribution of various hedonic and utilitarian elements based on the number of likes and comments, respectively (Table 3 ). First, the results suggest that seasonality did play a role in how users engage with the pictorial content. In both cases, there was a significant difference between winter (winter = 1) and summer (winter = 0), p likes <.001 and p comments <.001, in which, pictures posted in winter usually received higher engagement rate potentially due to several holidays during the season in Europe such as Christmas and New Year. Overall, the findings reveal that when pictures connoted a sense of thrill, there was a positive and significant difference in users’ liking and commenting behaviour, p likes <.001 and p comments =.001. Yet, pictures having functional positions had a significant and negative impact on the number of likes and comments, p likes <.001 and p comments <.001. Interestingly, while embedding fun elements in pictures significantly encouraged one’s liking behaviour, p < .001, it was ineffective in triggering a deeper level of engagement as reflected by the number of comments, p = 0.02. Furthermore, the results suggest that delivering a sense of usefulness, sensibility, and playfulness had a significant and negative effect on users’ intention to comment on a post. Due to the diversity of pictorial content, the subsequent analysis delves deeper into the varied effects of hedonic and utilitarian elements in different settings. Table 3 Summary of regression models without considering topics Liking behaviour Commenting behaviour coef. z-score coef. z-score Intercept -4.3930*** -462.79 -9.1170*** -737.078 Functional -0.0499*** -6.593 -0.0503*** -5.107 Necessary 0.0004 0.049 -0.0053 -0.554 Useful -0.0082 -1.071 -0.0431*** -4.344 Sensible -0.0092 -1.208 -0.0273** -2.757 Productive -0.0745*** -9.322 0.0074 0.712 Fun 0.0351*** 4.590 -0.0232* -2.328 Delightful -0.0058 -0.753 0.0127 1.255 Thrilling 0.0396*** 5 .088 0.0345** 3.398 Playful -0.0147 -1.882 -0.0350** -3.453 Cheerful -0.0177* -2.293 -0.0270* -2.677 Sensuous -0.0183** -2.571 -0.0001 -0.012 Funny -0.0079 -1.038 0.0019 0.190 Season 0.1723*** 12.862 0.1550*** 8.887 Note : *p < .05; **p < .01; ***p < .001 4.4. Regression models: Model selection and interactions According to the initial exploration of the scale (Table 3 ), this study then focuses on the covariates that were significant. Table 4 outlines the results of model selection on users’ liking and commenting behaviours with a significant effect. As topics were used as control, the models explain how average predictions changed across topics when the bipolar coordinate was held constant. Notably, Topic 0 served as the baseline. Thus, when every bipolar dimension was 0 and the Topic was 0, the intercept inferred the average engagement. The coefficients for all other topics were always relative to this baseline. That is, if they were positive, they were associated with an increase in the average engagement as compared to Topic 0, and vice versa. The findings indicate that most of the topics significantly influenced the number of likes in a positive way, except Topic 1 (Art and sculpture), 5 (Dining ambience), and 13 (Carriage horse rides). As for commenting behaviour, significant predictions were observed mostly in topics related to transportation [Topic 4 (Boat tours), 8 (Railway exploration), and 11 (Elevated urban transit)], in addition to Topic 5 (Dining ambience) and 7 (Church and architecture). Table 4 Model selection for liking and commenting behaviour Liking behaviour Commenting behaviour coef. z-score coef. z-score Intercept -4.4580*** -248.136 Intercept -9.1615*** -374.13 Topic 1 -0.0124 -0.498 Topic 1 0.0010 0.030 Topic 2 -0.0765** -2.882 Topic 2 -0.0670 -1.860 Topic 3 0.0820** 3.000 Topic 3 0.0213 0.574 Topic 4 0.1819*** 6.727 Topic 4 0.0758* 2.051 Topic 5 0.0050 0.167 Topic 5 0.2866*** 7.069 Topic 6 0.0810** 2.555 Topic 6 -0.0061 -0.141 Topic 7 0.1668*** 5.104 Topic 7 0.1240** 2.784 Topic 8 0.3097*** 7.963 Topic 8 0.1469** 2.783 Topic 9 0.2240*** 5.546 Topic 9 0.0981 1.787 Topic 10 0.1120** 2.723 Topic 10 0.0117 0.209 Topic 11 0.2563*** 6.058 Topic 11 0.1594** 2.770 Topic 12 0.1716*** 3.898 Topic 12 0.0645 1.076 Topic 13 0.0636 1.306 Topic 13 0.0464 0.699 Topic 14 0.2130** 3.083 Topic 14 0.1307 1.396 Note : *p < .05; **p < .01; ***p < .001 Further exploration involved fitting the models with topic interactions. While the majority of interactions did not exhibit significant differences, the noteworthy insights emerged from those interactions that did prove to be significant (Table 5 ), for functionality, usefulness, productiveness, thrill, and fun. First, in terms of the extent of functionality presenting in pictures, both Topic 5 (Dining ambience) and 10 (Cycling adventures) showed a negative impact on the number of likes and comments. Yet, there was a significant and positive relationship in one’s probability to like artistic pictures (Topic 1). Regarding pictures embedded thrilling experiences, markedly, the number of likes and comments increased alongside the level of thrill in Topic 10 (cycling adventures). Conversely, the results demonstrate significant but negative relations with both liking and commenting behaviour in pictures featuring riverside charm (Topic 6) and railway exploration (Topic 8). Likewise, when dining ambience (Topic 5) contained a higher extent of thrill, it created adverse effect on users’ intention to comment. Turning to the dimension of usefulness, interestingly, significant and negative impacts were found in all cases. In general, pictures featuring Ferris wheel and festivals (Topic 9) received less likes and comments in relation to the extent of usefulness. Similar patterns were observed in Topic 1 (Art and sculpture), 2 (Floral impressions) and 5 (Dining ambience) on commenting behaviour, and Topic 10 (Cycling adventures) on liking behaviour. Concerning the level of productivity, this study uncovers that connoting pictures with the quality of being productive encouraged one’s intention to like and comment on Topic 14 (Cityscape and tranquil scenery). This strategy was also effective in increasing the number of likes in pictures featuring floral impressions (Topic 2), railway exploration (Topic 8), and carriage horse rides (Topic 13). As for commenting behaviour, while embedding productive consumption in the pictures was effective for Topic 5 (Dining ambience), it decreased behavioural intention in Topic 7 (Church and architecture). Moreover, to the extent that pictures connotated fun consumption, the models suggest some unexpected but thought-provoking observations. Particularly, for all topics listed under multimodal excursions (Topic 4, 8, 10, 11, and 13), the degree of fun correlated negatively with the number of likes. Similar patterns were also found in Topic 5 (Dining ambience), 9 (Ferris wheel and festivals), 12 (Castle architecture), and 3 (Nature and parks). Finally, when sensibility was analysed with the topic, no significant interaction effect was discovered. Table 5 Topic interactions with significant effect Liking behaviour*Topic Commenting behaviour*Topic coef. z-score coef. z-score Functional -0.0216 -1.247 Functional -0.0388 -1.631 Topic 5 -0.0875** -2.881 Topic 1 0.0853** 2.330 Topic 10 -0.1655** -3.303 Topic 5 -0.1006** -2.429 Topic 10 -0.1659** -2.441 Useful -0.0008 -0.049 Useful 0.006. 0.268 Topic 9 -0.0887* -2.010 Topic 1 -0.0992** -3.032 Topic 10 -0.1527** -2.712 Topic 2 -0.0844* -2.251 Topic 5 -0.0764* -2.114 Topic 9 -0.1234* -2.155 Productive -0.0943*** -5.526 Productive -0.0206 -0.891 Topic 2 0.0637* 2.317 Topic 5 0.0792* 2.242 Topic 8 0.1298* 2.588 Topic 7 -0.1722** -2.667 Topic 13 0.1475* 2.408 Topic 14 0.5254** 2.978 Topic 14 0.4926*** 4.064 Thrilling 0.0433** 2.598 Thrilling 0.0716** 3.183 Topic 6 -0.0989** -3.098 Topic 5 -0.1680*** -3.767 Topic 8 -0.1996*** -4.22 Topic 6 -0.0992* -2.355 Topic 10 0.1505** 2.731 Topic 8 -0.1672** -2.643 Topic 10 0.1391* 2.001 Fun 0.1062*** 7.542 Sensible -0.0124 -0.577 Topic 3 -0.0692** -2.708 Topic 4 -0.0803** -2.622 Topic 5 -0.0772** -3.036 Topic 8 -0.1912*** -4.521 Topic 9 -0.1954*** -5.535 Topic 10 -0.1287* -2.414 Topic 11 -0.2547*** -3.912 Topic 12 -0.1363* -2.202 Topic 13 -0.1116* -2.187 Note : *p < .05; **p < .01; ***p < .001 5. Discussion In a landscape where tourism experiences are conveyed visually (Xu et al., 2024), this research seeks to delve into the intricate relationship between the embedded hedonic and utilitarian aspects in images and their influence on user engagement. Nevertheless, it is important to recognise that this focus does not diminish the significance of other elements, such as post captions, popular hashtags, or alternative formats like videos. Notably, although this study centres on pictures, other additional components also play crucial roles in enhancing visibility, providing context, and boosting engagement (Fu et al., 2024 ; Yu et al., 2025 ). Interestingly, unlike the commonly identified destination image attributes such as nature and outdoor environment (e.g., beach, mountain, forest, lake), gastronomy, architectures (e.g., museum, heritage, temple), as well as cityscape summarised in existing literature (Picazo & Moreno-Gil, 2019 ; Yu & Egger, 2021 ), insights from image captions reveal a new facet on multimodal excursions. This includes boat rides, railway journeys, cycling, urban transit, and horse rides. These often-overlooked aspects contribute to a more comprehensive understanding of tourist experiences. However, in fact, the emphasis on the functional dimensions of tourism products, such as transportation modes, has been strengthened in research that analyses destination image through text mining based on online reviews (Li et al., 2022 ). This trend shows that traditional image attributes are extending to pictorial representations, highlighting that functional aspects, such as various transportation modes, are integral to overall travel experiences (Költringer & Dickinger, 2015 ). Subsequently, the influence of different topics on potential tourists’ liking and commenting behaviours is scrutinised. Concerning experiences related to transportation, these findings challenge preconceived notions regarding the universal appeal of fun in driving engagement (Vries et al., 2012 ). Despite tourism is about selling an emotional journey (Yu & Egger, 2021 ), this research initiates a reassessment of assumptions regarding user preferences and the impact of fun on shaping engagement across varied thematic contexts. Furthermore, although previous research found that incorporating activity-centric pictures can enhance user engagement (Aramendia-Muneta et al., 2020 ), it is important to consider the context. For multimodal excursions, users may unconsciously prefer informative content over entertaining appeals (Hong et al., 2023 ). While fun consumption was not prominent in multimodal excursions, thrill significantly increased user engagement in pictures of cycling adventures (Fossgard & Fredman, 2019 ). This heightened engagement likely stems from the dynamic nature of cycling, aligning with thrill-seeking expectations in adventure tourism (Kiatkawsin et al., 2021 ). This is further supported when considering functionality. A functional atmosphere in pictures might detract from desired experiences, discouraging engagement. Similarly, users might avoid engaging with serene visuals (e.g., riverside) if thrill elements are present, as they may prioritise tranquillity over excitement in nature-based tourism (Conti & Lexhagen, 2020 ). Furthermore, the negative impact on commenting behaviour with dining experiences adds complexity to thrill consumption. This could be that tourists often seek relaxed, immersive gastronomic experiences (Dixit & Prayag, 2022 ), and introducing thrilling elements might disrupt this atmosphere, discouraging active commenting. Regarding art and sculpture, the findings highlight the intricate interplay between functionality and thematic context. The perceived utility of art and sculpture can enhance user engagement rather than hinder it. When considering the multifaceted nature of art appreciation, functionality might be perceived as a means to deepen one’s understanding with the artistic elements (Liu, 2019 ). This integration acts as a bridge between aesthetic and practical dimensions (Botti, 2000 ), catering to diverse user preferences that find such content enriching and intellectually stimulating. Conversely, since art is often regarded as a form of expression and creativity (Botti, 2000 ), users may be less motivated to engage with content perceived as merely useful (Christiaans, 2002). These findings challenge the assumption that perceived usefulness always positively impacts user preferences (Aramendia-Muneta et al., 2020 ). In tourism marketing, which heavily relies on visual appeals (Arabadzhyan et al., 2021), user engagement dynamics may deviate from utility-based expectations (Chen & Fu, 2018 ). This aligns with the focus on experiential marketing, where tourists seek dynamic encounters that go beyond traditional notions of usefulness (Tussyadiah, 2014 ). Lastly, cityscape and tranquil scenery images highlight the productivity of visual content, prompting potential tourists to show appreciation through likes and comments. A potential explanation for this phenomenon is that individuals perceive a sense of productivity, akin to accomplishment (Leitão et al., 2021 ), when engaging with tranquil content. Such engagement, therefore, fosters a positive mental state, leading users to express their appreciation (Neuhofer et al., 2021 ). In another scenario, this study implies that users value content showcasing culinary productivity within dining ambience, such as the artistic and skilful aspects. This aligns with the prevailing trend of culinary exploration and visually appealing food experiences on social media (Gambetti & Han, 2022 ). Nonetheless, users exploring church and architecture content might prioritise cultural significance and aesthetic qualities of the depicted locations (López-Chao & Lopez-Pena, 2020 ). In these contemplative settings, productive consumption may seem less relevant or disruptive to the experience. 6. Conclusion Essentially, this research investigates the relationship between the hedonic and utilitarian aspects of tourism-related images shared on social media and their impact on user engagement. By applying a machine-generated semantic differential approach, it underscores that user engagement in tourism imagery is influenced by thematic context, challenging the assumption that fun universally drives engagement. For instance, functionality can enhance engagement with art and sculpture by deepening understanding, whereas perceived usefulness does not always correlate with engagement. Meanwhile, although thrill enhances engagement in cycling adventures, users may prefer informative content over entertaining appeals in multimodal excursions. Overall, the results provide a nuanced understanding of the interplay between emotional connotations in visual content and potential tourists’ engagement on social media. The following sections delve deeper into the theoretical and practical implications. 6.1. Theoretical contributions The interdisciplinary nature of this study makes theoretical advancements by bridging consumer psychology, data analytics, and emotional experience analysis, within the context of tourism visual marketing. Specifically, one of the key contributions lies in the introduction of a novel methodological approach that applies machine-generated semantic differentials to evaluate the interplay between hedonic and utilitarian values embedded in visual content. Given the inherent semantic variations in images, the application of machine learning in this study provides unique insights for future research. Distinct from traditional survey-based methods, one key advantage of the bipolar framework is its objectivity in measuring experiences (Brühlmann et al., 2020 ). By elevating the subjective aspects of semantic differentials to an objective level, the machine-generated approach enhances the robustness of the results. By transcending traditional approaches, this technique addresses the gap left by the lack of focus on emotional experiences in the analysis of tourism pictures (Vu et al., 2025 ), highlighting the need for more comprehensive exploration. For instance, the dimension of usefulness challenges traditional assumptions, revealing that tourists in entertainment-oriented themes prioritise emotional and experiential content over practical information. The results also stimulate discussions regarding the broad popularity of enjoyment across different settings (Vries et al., 2012 ), prompting a re-evaluation of user preferences within different tourism contexts. Moreover, the use of data-driven techniques (Mathew et al., 2020 ) extends the status quo of the hedonic-utilitarian measurement (Voss et al., 2003 ). By shedding light on the importance of hedonic and utilitarian consumption in social media marketing, this research emphasises the need to create visually appealing content that evokes various feelings, depending on the context, to attract and engage potential tourists (Gambetti & Han, 2022 ). Additionally, beyond conventional themes commonly explored in previous literature, such as nature, dining, architecture, and cityscape (Picazo & Moreno-Gil, 2019 ), the identification of multimodal excursions (e.g., boat, railway, and horse rides) adds a new dimension to the visual representation of tourism experiences on social media. The utilisation of image captions for labelling also introduces an emerging approach in visual analysis research (He et al., 2020 ) as it potentially reveals underexplored topics in destination pictures. This finding challenges the existing paradigms that predominantly focus on traditional attributes such as nature and gastronomy (Conti & Lexhagen, 2020 ; Picazo & Moreno-Gil, 2019 ; Yu & Egger, 2021 ). By highlighting the emotional and experiential dimensions associated with these often-overlooked themes, this study encourages future research to explore the complexities of tourism experiences beyond conventional boundaries. Moreover, the findings encourage a re-evaluation of the hedonic-utilitarian dichotomy in consumption literature. They prompt scholars to explore the contextual factors influencing behaviour, enriching the theoretical discourse on emotional consumption in tourism and digital marketing. Overall, this research paves the way for the integration of new techniques beyond tourism, into other services and marketing disciplines, emphasising the importance of understanding and catering to diverse consumer preferences within specific thematic contexts for more effective and resonant strategies. 6.2. Practical implications By establishing a vital connection between visual elements and the efficacy of embedding diverse consumption experiences, this study offers destination marketers valuable insights for optimising social media content and engaging tourists effectively. For instance, one of the findings indicates that content centred around multimodal excursions tends to prioritise informativeness over entertainment. That is, marketers should focus on crafting visually engaging posts that emphasise the practical aspects of transportation experiences—such as safety, efficiency, and accessibility—while still conveying emotional resonance. In the realm of dining experiences, the study reveals a relationship between functionality and engagement. The results imply that marketers should consider integrating aspects of artistry and skill into their culinary visuals, as these attributes can foster deeper emotional connections. For example, posts that showcase chefs in action or beautifully plated dishes can enhance user appreciation and interaction. The study also emphasises the positive response to thrill in adventurous activities (e.g., cycling), indicating alignment with expectations in adventure tourism. Hence, content creators can leverage this insight by showcasing dynamic imagery that captures the excitement of cycling or boat tours. For instance, an Instagram post featuring a cyclist navigating through a vibrant landscape, paired with exhilarating captions, can effectively resonate with thrill-seekers and encourage higher interaction rates. However, marketers should exercise caution to avoid disrupting desired atmospheres, especially in serene environments. Moreover, exploring the merging of functionality with artistic expression provides another avenue for marketers. This trend reflects a preference for enhanced aesthetic experiences and interactive engagement, which is in line seamlessly with the contemporary emphasis on experiential marketing. In a nutshell, the insights from this research emphasise the necessity for marketers to adopt a tailored approach to their visual content strategies, recognising that the impact of hedonic and utilitarian elements varies significantly across different tourism contexts. Recognising these intricacies is crucial for destination marketers seeking to enhance user engagement and the overall effectiveness in visual marketing. 6.3. Limitations and recommendations While this study provides valuable insights, it is not without its limitations. Firstly, despite the novelty of the data-driven bipolar framework, the study primarily focuses on the hedonic and utilitarian dimensions in isolation, potentially overlooking the full richness of tourism experiences. Future research is recommended to explore alternative measurement methods in experiential marketing in order to broaden the understanding of the impact of visual content on tourism. Researchers may also investigate the interactive effects and potential conflicts between these dimensions for a more comprehensive view. Furthermore, although efforts were made to ensure data source diversity, the focus on European destinations may fall short of capturing the unique cultures or atmospheres found in other regions. Additionally, since this study is conducted from the perspective of destination marketers, it is important to recognise that their viewpoint might differ from that of ordinary tourists. To address this limitation, future scholars should consider expanding the scope to include a more diverse range of destinations across various continents. Comparing these findings with tourists’ perceived images could provide further insights. Meanwhile, in addition to analysing likes and comments, scholars are encouraged to delve deeper by examining the content of comments. This can be achieved through sentiment or emotional analysis to quantify user reactions, or by employing qualitative methods to gain richer insights. Finally, with the rise of short videos in the tourism domain, future studies are encouraged to delve into the distinctive attributes and effects of video content on user engagement. Additionally, factors such as post captions should be examined in tandem. This exploration would contribute to a more comprehensive understanding of the evolving landscape of visual content in marketing. Declarations Author Contribution Joanne Yu: Conceptualization, Writing - Original Draft, Writing - Review & Editing; Denis Helic: Methodology, Analysis, Writing - Review & Editing; Astrid Dickinger: Conceptualization, Writing - Review & Editing; Markus Strohmaier: Methodology, Analysis References Aramendia-Muneta ME, Olarte-Pascual C, Ollo-López A (2020) Key Image Attributes to Elicit Likes and Comments on Instagram. J Promotion Manage 1–27. https://doi.org/10.1080/10496491.2020.1809594 Arefieva V, Egger R, Yu J (2021) A machine learning approach to cluster destination image on Instagram. Tour Manag 85:104318. https://doi.org/10.1016/j.tourman.2021.104318 Botti S (2000) What Role for Marketing in the Arts? An Analysis of Arts Consumption and Artistic Value. Int J Arts Manage, 14–27 Brühlmann F, Petralito S, Rieser DC, Aeschbach LF, Opwis K (2020) TrustDiff: Development and Validation of a Semantic Differential for User Trust on the Web. J Usability Stud 16(1):29–48 Chen JH, Fu J-R (2018) On the effects of perceived value in the mobile moment. Electron Commer Res Appl 27:118–128. https://doi.org/10.1016/j.elerap.2017.12.009 City Destinations Alliance (2022) CityDNA Benchmarking Report . https://citydestinationsalliance.eu/citydna-benchmarking-report/ Conti E, Lexhagen M (2020) Instagramming nature-based tourism experiences: a netnographic study of online photography and value creation. Tourism Manage Perspect 34:100650. https://doi.org/10.1016/j.tmp.2020.100650 Das G, Varshneya G (2017) Consumer emotions: Determinants and outcomes in a shopping mall. J Retailing Consumer Serv 38:177–185. https://doi.org/10.1016/j.jretconser.2017.06.008 Davari D, Nosrati S, Kim S (2024) Do cultural and individual values influence sustainable tourism and pro-environmental behavior? Focusing on Chinese millennials. J Travel Tourism Mark 41(4):559–577. https://doi.org/10.1080/10548408.2024.2309180 Deng N, Li X (2018) Feeling a destination through the right photos: A machine learning model for DMOs’ photo selection. Tour Manag 65:267–278. https://doi.org/10.1016/j.tourman.2017.09.010 Dixit SK, Prayag G (2022) Gastronomic tourism experiences and experiential marketing. Tourism Recreation Res 47(3):217–220. https://doi.org/10.1080/02508281.2022.2065089 Engler J, Sikdar S, Lutz M, Strohmaier M (2023) SensePOLAR: Word sense aware interpretability for pre-trained contextual word embeddings. ArXiv Preprint ArXiv:2301.04704 Fellbaum C (2005) WordNet and wordnets. In: Brown, Keith (eds.), Encyclopedia of Language and Linguistics , Second Edition, Oxford: Elsevier, 665–670 Filieri R, Yen DA, Yu Q (2021) #ILoveLondon: An exploration of the declaration of love towards a destination on Instagram. Tour Manag 85:104291. https://doi.org/10.1016/j.tourman.2021.104291 Fossgard K, Fredman P (2019) Dimensions in the nature-based tourism experiencescape: An explorative analysis. J Outdoor Recreation Tourism 28:100219. https://doi.org/10.1016/j.jort.2019.04.001 Fu X, Liu X, Li Z (2024) Catching eyes of social media wanderers: How pictorial and textual cues in visitor-generated content shape users’ cognitive-affective psychology. Tour Manag 100:104815 Gambetti A, Han Q (2022) Camera eats first: exploring food aesthetics portrayed on social media using deep learning. Int J Contemp Hospitality Manage 34(9):3300–3331. https://doi.org/10.1108/IJCHM-09-2021-1206 Gon M (2020) Local experiences on Instagram: Social media data as source of evidence for experience design. J Destination Mark Manage 100435. https://doi.org/10.1016/j.jdmm.2020.100435 Grootendorst M (2020) Bertopic: Leveraging bert and c-tf-idf to create easily interpretable topics. Version V0, 4 Hauser D, Leopold A, Egger R, Ganewita H, Herrgessell L (2022) Aesthetic perception analysis of destination pictures using #beautifuldestinations on Instagram. J Destination Mark Manage 24:100702. https://doi.org/10.1016/j.jdmm.2022.100702 He S, Liao W, Tavakoli HR, Yang M, Rosenhahn B, Pugeault N (2020) Image Captioning through Image Transformer. Proceedings of the Asian Conference on Computer Vision , 1–17 Hong WCH, Ngan HFB, Yu J, Arbouw P (2023) Examining cultural differences in Airbnb naming convention and user reception: an eye-tracking study. J Travel Tourism Mark 40(6):475–489. https://doi.org/10.1080/10548408.2023.2263764 Hossain MZ, Sohel F, Shiratuddin MF, Laga H (2019) A comprehensive survey of deep learning for image captioning. ACM Comput Surv (CsUR) 51(6):1–36 Huang X, Kader MS, Kim S (2021) Let's play with emojis! How to make emojis more effective in social media advertising using promocodes and temporal orientation. Journal Res Interact Mark Adv online publication. https://doi.org/10.1108/JRIM-04-2021-0113 Huang Y, Chen CH, Khoo LP (2012) Products classification in emotional design using a basic-emotion based semantic differential method. Int J Ind Ergon 42(6):569–580. https://doi.org/10.1016/j.ergon.2012.09.002 Kiatkawsin K, Bui NA, Hrankai R, Jeong K (2021) The Moderating Roles of Sensation Seeking and Worry among Nature-Based Adventure Tourists. Int J Environ Res Public Health 18(4). https://doi.org/10.3390/ijerph18042021 Költringer C, Dickinger A (2015) Analyzing destination branding and image from online sources: A web content mining approach. J Bus Res 68(9):1836–1843 Kuo L, Chang T, Lai C–C (2022) Multimedia webpage visual design and color emotion test. Multimedia Tools Appl 81(2):2621–2636. https://doi.org/10.1007/s11042-021-11684-4 Lee Y–K, Lee C–K, Lee W, Ahmad MS (2021) Do hedonic and utilitarian values increase pro-environmental behavior and support for festivals? Asia Pac J Tourism Res 26(8):921–934. https://doi.org/10.1080/10941665.2021.1927122 Leitão J, Pereira D, Gonçalves  (2021) Quality of Work Life and Contribution to Productivity: Assessing the Moderator Effects of Burnout Syndrome. Int J Environ Res Public Health 18(5). https://doi.org/10.3390/ijerph18052425 Li H, Ji H, Liu H, Cai D, Gao H (2022) Is a picture worth a thousand words? Understanding the role of review photo sentiment and text-photo sentiment disparity using deep learning algorithms. Tour Manag 92:104559. https://doi.org/10.1016/j.tourman.2022.104559 Li J, Abbasi A, Cheema A, Abraham LB (2020) Path to Purpose? How Online Customer Journeys Differ for Hedonic Versus Utilitarian Purchases. J Mark 84(4):127–146. https://doi.org/10.1177/0022242920911628 Li Y, Xie Y (2020) Is a Picture Worth a Thousand Words? An Empirical Study of Image Content and Social Media Engagement. J Mark Res 57(1):1–19. https://doi.org/10.1177/0022243719881113 Liadeli G, Sotgiu F, Verlegh PW (2023) A Meta-Analysis of the Effects of Brands’ Owned Social Media on Social Media Engagement and Sales. J Mark 87(3):406–427. https://doi.org/10.1177/00222429221123250 Lim H, Childs M (2020) Visual storytelling on Instagram: branded photo narrative and the role of telepresence. J Res Interact Mark 14(1):33–50. https://doi.org/10.1108/JRIM-09-2018-0115 Lim JS, Choe M–J, Zhang J, Noh G–Y (2020) The role of wishful identification, emotional engagement, and parasocial relationships in repeated viewing of live-streaming games: A social cognitive theory perspective. Comput Hum Behav 108:106327. https://doi.org/10.1016/j.chb.2020.106327 Lin F, Ryu K, Lee YK, Ali F (2024) Understanding the importance of café portable cup design and environmental concerns on green on-the-go consumption: a comparative study of gender. J Travel Tourism Mark 41(9):1262–1279. https://doi.org/10.1080/10548408.2024.2425101 Liu Q, Kusner MJ, Blunsom P (2020) A survey on contextual embeddings. arXiv preprint arXiv:2003.07278 Liu M (2019) Correlation between Aesthetic Ability and Art Appreciation. Int J Trend Res Dev 6(5):14–16 López-Chao V, Lopez-Pena V (2020) Aesthetical Appeal and Dissemination of Architectural Heritage Photographs in Instagram. Buildings 10(12):225. https://doi.org/10.3390/buildings10120225 Ma Z, Palacios S (2021) Image-mining: exploring the impact of video content on the success of crowdfunding. J Mark Analytics 9(4):265–285. https://doi.org/10.1057/s41270-021-00133-8 Mathew B, Sikdar S, Lemmerich F, Strohmaier M (2020) The POLAR Framework: Polar Opposites Enable Interpretability of Pre-Trained Word Embeddings. In Y. Huang, I. King, T.-Y. Liu, & M. van Steen (Eds.), Proceedings of The Web Conference 2020 (pp. 1548–1558). ACM. https://doi.org/10.1145/3366423.3380227 McInnes L, Healy J, Saul N, Großberger L (2018) UMAP: Uniform Manifold Approximation and Projection. J Open Source Softw 3(29):861. https://doi.org/10.21105/joss.00861 Mindak WA (1961) Fitting the semantic differential to the marketing problem. J Mark 25(4):28–33 Neuhofer B, Egger R, Yu J, Celuch K (2021) Designing experiences in the age of human transformation: An analysis of Burning Man. Annals Tourism Res 91:103310. https://doi.org/10.1016/j.annals.2021.103310 Nguyen TK, Coustaty M, Guillaume JL (eds) (2018) A new image segmentation approach based on the Louvain algorithm . IEEE. http://ieeexplore.ieee.org/servlet/opac?punumber=8491725 Picazo P, Moreno-Gil S (2019) Analysis of the projected image of tourism destinations on photographs: A literature review to prepare for the future. J Vacation Mark 25(1):3–24. https://doi.org/10.1177/1356766717736350 Radford A, Kim JW, Hallacy C, Ramesh A, Goh G, Agarwal S, Sastry G, Askell A, Mishkin P, Clark J, Krueger G, Sutskever I (2021), February 26 Learning Transferable Visual Models From Natural Language Supervision . http://arxiv.org/pdf/2103.00020v1 Reelfs J, Mohaupt T, Sikdar S, Strohmaier M, Hohlfeld O (eds) (2022) Interpreting Emoji with Emoji. Association for Computational Linguistics Reimers N, Gurevych I (2019) Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP) , 3982–3992, Association for Computational Linguistics Rietveld R, van Dolen W, Mazloom M, Worring M (2020) What You Feel, Is What You Like Influence of Message Appeals on Customer Engagement on Instagram. J Interact Mark 49:20–53. https://doi.org/10.1016/j.intmar.2019.06.003 Roggeveen AL, Grewal D, Townsend C, Krishnan R (2015) The Impact of Dynamic Presentation Format on Consumer Preferences for Hedonic Products and Services. J Mark 79(6):34–49. https://doi.org/10.1509/jm.13.0521 Soldat Z, Quintal V, Phau I (2024) Pursuing travel calculated hedonism: scale development and validation. J Travel Tourism Mark 41(5):760–774. https://doi.org/10.1080/10548408.2024.2341653 Stead S, Wetzels R, Wetzels M, Odekerken-Schröder G, Mahr D (2022) Toward Multisensory Customer Experiences: A Cross-Disciplinary Bibliometric Review and Future Research Directions. J Service Res 109467052210799. https://doi.org/10.1177/10946705221079941 Stremersch S, Gonzalez J, Valenti A, Villanueva J (2023) The value of context-specific studies for marketing. J Acad Mark Sci 51(1):50–65. https://doi.org/10.1007/s11747-022-00872-9 Su L, Ye C, Huang Y (2024) Does destination nostalgic advertising enhance tourists’ intentions to visit? The moderating role of destination type. Tour Manag 100:104810. https://doi.org/10.1016/j.tourman.2023.104810 Sung YH, Kim DH, Choi D, Lee SY (2020) Facebook ads not working in the same way: The effect of cultural orientation and message construals on consumer response to social media ads. Telematics Inform 52:101427. https://doi.org/10.1016/j.tele.2020.101427 Taberner I, Juncà A (2021) Small-Scale Sport Events as Place Branding Platforms: A Content Analysis of Osona’s Projected Destination Image through Event-Related Pictures on Instagram. Sustainability 13(21):12255. https://doi.org/10.3390/su132112255 Taecharungroj V, Mathayomchan B (2021) Traveller-generated destination image: Analysing Flickr photos of 193 countries worldwide. Int J Tourism Res 23(3):417–441 Tussyadiah IP (2014) Toward a Theoretical Foundation for Experience Design in Tourism. J Travel Res 53(5):543–564. https://doi.org/10.1177/0047287513513172 Villamediana-Pedrosa JD, Vila-Lopez N, Küster-Boluda I (2018) Secrets to design an effective message on Facebook: an application to a touristic destination based on big data analysis. Curr Issues Tourism 13(6):1–21. https://doi.org/10.1080/13683500.2018.1554625 Virtanen H, Björk P, Sjöström E (2017) Follow for follow: marketing of a start-up company on Instagram. J Small Bus Enterp Dev 24(3):468–484 Villamediana J, Küster I, Vila N (2019) Destination engagement on Facebook: Time and seasonality. Annals Tourism Res 79:102747 Voss KE, Spangenberg ER, Grohmann B (2003) Measuring the hedonic and utilitarian dimensions of consumer attitude. J Mark Res 40(3):310–320 de Vries L, Gensler S, Leeflang PS (2012) Popularity of Brand Posts on Brand Fan Pages: An Investigation of the Effects of Social Media Marketing. J Interact Mark 26(2):83–91. https://doi.org/10.1016/j.intmar.2012.01.003 Vu HQ, Song B, Li G, Law R (2025) Exploring emotional aspects of travel concepts via travel photos based on contrastive language-image pretraining. Tour Manag 108:105117 Wang L, Luo Z (2025) Effects of tourists’ online self-disclosure depth on readers’ visit intentions: mediating roles of perceived social distance and perceived self-expressiveness. J Travel Tourism Mark 42(1):40–64. https://doi.org/10.1080/10548408.2024.2425100 Wies S, Bleier A, Edeling A (2023) Finding Goldilocks Influencers: How Follower Count Drives Social Media Engagement. J Mark 87(3):383–405. https://doi.org/10.1177/00222429221125131 Wu L, You JS (2022) Synchronous or Asynchronous Course: Business Students’ Perspectives on an Optimized Modality of Online Teaching and Learning. J Mark Educ 44(2):265–284. https://doi.org/10.1177/02734753221093740 Xiao X, Fang C, Lin H (2020) Characterizing Tourism Destination Image Using Photos’ Visual Content. ISPRS Int J Geo-Information 9(12):730. https://doi.org/10.3390/ijgi9120730 Xu J, Wang Y, Jiang Y (2023) How do social media tourist images influence destination attitudes? Effects of social comparison and envy. J Travel Tourism Mark 40(4):310–325. https://doi.org/10.1080/10548408.2023.2245410 You Q, Jin H, Wang Z, Fang C, Luo J (2016) Image captioning with semantic attention. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 4651–4659) Yu J, Egger R (2021) Color and engagement in touristic Instagram pictures: A machine learning approach. Annals Tourism Res. 103204 https://doi.org/10.1016/j.annals.2021.103204 Yu J, Hong WCH, Egger R (2025) The art of post captions: Readability and user engagement on social media. J Travel Res 64(4):853–866 Yuda Bakti IGM, Rakhmawati T, Sumaedi S, Damayanti S (2020) Railway commuter line passengers’ perceived service quality: hedonic and utilitarian framework. Transp Res Procedia 48:207–217. https://doi.org/10.1016/j.trpro.2020.08.016 Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 29 Apr, 2026 Read the published version in Information Technology & Tourism → Version 1 posted Editorial decision: Revision requested 10 Nov, 2025 Reviews received at journal 06 Nov, 2025 Reviews received at journal 05 Nov, 2025 Reviewers agreed at journal 08 Oct, 2025 Reviewers agreed at journal 06 Oct, 2025 Reviewers invited by journal 01 Oct, 2025 Editor assigned by journal 29 Sep, 2025 Submission checks completed at journal 15 Sep, 2025 First submitted to journal 11 Sep, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7591283","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":528223463,"identity":"0279520a-b1ad-463f-8d58-fc5e0f66c1db","order_by":0,"name":"Joanne Yu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA40lEQVRIiWNgGAWjYBACxgYogx/OPECsFskGxsYGorTAgcEBmHZCWpjbm589Lqg4nLj5+OH2x7xtDHJ8NxJYN3zA57CeY+bGM84cTtx2JrGxGajFWPJGAtvNGfi0zEgwk+ZtA2q5wQjWkrgBqOU2D14t6d+kef8BHTYDoqWeCC05QFsaDidukIBoSTAgqKXnTJk0z7F0oH8SG2fOOSdhOPPMwza8fjFsb98mzVNjLdvffvzBhzdlNvJ8x5OP3cAXYoYNYKoZTDLxMEgwIEUvdiAPoeogrvyBV+0oGAWjYBSMVAAARH1VZIQPjmUAAAAASUVORK5CYII=","orcid":"","institution":"City University of Macau","correspondingAuthor":true,"prefix":"","firstName":"Joanne","middleName":"","lastName":"Yu","suffix":""},{"id":528223464,"identity":"8286cee9-4094-485d-88d3-42a5cdf3f7c4","order_by":1,"name":"Denis Helic","email":"","orcid":"","institution":"Graz University of Technology","correspondingAuthor":false,"prefix":"","firstName":"Denis","middleName":"","lastName":"Helic","suffix":""},{"id":528223465,"identity":"0df1279d-ddb6-433a-a220-99e02899d595","order_by":2,"name":"Astrid Dickinger","email":"","orcid":"","institution":"MODUL University Vienna","correspondingAuthor":false,"prefix":"","firstName":"Astrid","middleName":"","lastName":"Dickinger","suffix":""},{"id":528223466,"identity":"a4281bfa-2121-4893-bf19-a87d15d434f4","order_by":3,"name":"Markus Strohmaier","email":"","orcid":"","institution":"University of Mannheim","correspondingAuthor":false,"prefix":"","firstName":"Markus","middleName":"","lastName":"Strohmaier","suffix":""}],"badges":[],"createdAt":"2025-09-11 11:23:14","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7591283/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7591283/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s40558-026-00377-z","type":"published","date":"2026-04-29T15:58:03+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":93566328,"identity":"0667fd3b-d969-46e5-ac5b-beb4998b5252","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4481415,"visible":true,"origin":"","legend":"","description":"","filename":"ITTPOLARManuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/050a7b8c01859de9fc514180.docx"},{"id":93567153,"identity":"ca17cb59-113d-4aa4-aaaa-f16732f5bf73","added_by":"auto","created_at":"2025-10-15 08:38:45","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5342,"visible":true,"origin":"","legend":"","description":"","filename":"9c11c1c35f2e4d9c97417515b1197f1d.json","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/d8217ef0d099e32addffca65.json"},{"id":93566320,"identity":"5a77696a-7061-4700-bbad-8efd18b5b7fe","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":212590,"visible":true,"origin":"","legend":"","description":"","filename":"9c11c1c35f2e4d9c97417515b1197f1d1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/c68d0c20120174f5e4cc243d.xml"},{"id":93566315,"identity":"d57ac1ce-8e5b-49cb-8bfc-d8b987abeb48","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"png","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":272070,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/fe483b0eb8f2c2bafa2e1b3c.png"},{"id":93566317,"identity":"8fd3c3cf-6c53-4996-b38e-351ea0201690","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":225026,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/b7999dea7b829964b9e5313a.png"},{"id":93566333,"identity":"a0ac098f-ad89-4679-bde1-60283e14e073","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":249017,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/cbc5551f65d6f4dcc1577075.png"},{"id":93566335,"identity":"8608496c-8390-4730-a25a-995227f655ce","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":255317,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage12.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/8c51811211cdd8afa529c76b.png"},{"id":93567158,"identity":"b9fbbdd5-88bc-4b12-95d7-6cd26f33486a","added_by":"auto","created_at":"2025-10-15 08:38:46","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":259604,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage13.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/250abd7584a2b7ab396fce57.png"},{"id":93566329,"identity":"41d5b6c9-b308-49a0-a698-6b0c2c989369","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":240782,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage14.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/60eb9621a081807ea9ecc7f5.png"},{"id":93566344,"identity":"cdcd2393-2731-45ec-94cb-bd34779fca2b","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":239268,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage15.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/8c2a98f80446bb08883607e3.png"},{"id":93567169,"identity":"8288d1b8-902a-40e4-9200-8e71e3878a10","added_by":"auto","created_at":"2025-10-15 08:38:47","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":256064,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage16.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/a45f34ad1c0f3abc11ff5096.png"},{"id":93567155,"identity":"2009fac9-e72e-4a54-8b65-397cf291db9c","added_by":"auto","created_at":"2025-10-15 08:38:46","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":85529,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage17.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/3ca0021322c4eca07672aea2.png"},{"id":93566353,"identity":"10ca0321-4a6f-4cb9-905e-abf20bdbf960","added_by":"auto","created_at":"2025-10-15 08:30:47","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":88618,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage18.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/944114e1bddd13fea259dcab.png"},{"id":93568197,"identity":"26270a67-6962-49f3-b37c-bb2b7453ca57","added_by":"auto","created_at":"2025-10-15 08:46:47","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":86135,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage19.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/31e0260a0e97e3ae59f2990c.png"},{"id":93566355,"identity":"d04c68e8-b216-4411-bf3c-7152e4b15552","added_by":"auto","created_at":"2025-10-15 08:30:47","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":253294,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/3a86ac0cc9cdb82ed088ee79.png"},{"id":93567159,"identity":"e9f1f44f-8380-4c3d-9abf-15494a6fa9e6","added_by":"auto","created_at":"2025-10-15 08:38:46","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":253816,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/7b24d5f8a0ade83f76e9d580.png"},{"id":93566327,"identity":"160b67d6-f466-4f10-97bd-e292ee171fd0","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":229699,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/69637d08739cc225dca573e1.png"},{"id":93566322,"identity":"5fcd924c-aa6f-4fe2-8c8c-c65c053348bf","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":223341,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/ec7afa8fc2614a5c98df0172.png"},{"id":93567163,"identity":"45e626b0-31e1-4df3-8da4-a1e754416185","added_by":"auto","created_at":"2025-10-15 08:38:46","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":247121,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/5db8d2293d1c3d96ebf40f00.png"},{"id":93567177,"identity":"f1889bc7-eba6-4bb2-8133-1e060457b2e9","added_by":"auto","created_at":"2025-10-15 08:38:47","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":260419,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/d44c007373532759a5655105.png"},{"id":93566342,"identity":"8dadc3be-bc27-4683-b854-77a309fdf2d2","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":237680,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/8b89e164c2cc4379b8cb8829.png"},{"id":93566351,"identity":"46563f6d-a0ff-4186-867b-0ac52d93996c","added_by":"auto","created_at":"2025-10-15 08:30:47","extension":"png","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":255050,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/020e9737b0b8f1fcb232a412.png"},{"id":93566325,"identity":"58d685c3-15ae-4f7e-8a1d-dcdd59cc7cbf","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"png","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":184045,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/36e21776ecacf126da1335f3.png"},{"id":93566339,"identity":"fef1f28b-b641-40bb-a9fd-47a9b440a4a9","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"png","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":30319,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/5f1d2b40f77fb3f087c3b44d.png"},{"id":93566332,"identity":"6d1d17cc-2e96-4840-98a8-7be715bbbff5","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"png","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":36917,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/e81063cc1930f806a08f8dd1.png"},{"id":93567165,"identity":"8b53cd79-b932-49a2-bc67-92d97930b082","added_by":"auto","created_at":"2025-10-15 08:38:46","extension":"png","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":44395,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage12.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/2f42e188ab928b69c49c5109.png"},{"id":93567176,"identity":"372c4d0c-b083-4204-9841-4cdaacf75cb5","added_by":"auto","created_at":"2025-10-15 08:38:47","extension":"png","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":42052,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage13.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/3cffb2d0f0cf68731c840722.png"},{"id":93566341,"identity":"6bb05381-d932-458b-92ec-3423a6f6f14f","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"png","order_by":27,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":38379,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage14.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/f0917fed12b2860e56436c33.png"},{"id":93566349,"identity":"14f96795-a9dd-460d-805f-cad79e4e2d8d","added_by":"auto","created_at":"2025-10-15 08:30:47","extension":"png","order_by":28,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":40125,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage15.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/749cfec6e1629324ce91e84b.png"},{"id":93567175,"identity":"046b0681-918d-4124-aa7c-87011c71e44d","added_by":"auto","created_at":"2025-10-15 08:38:47","extension":"png","order_by":29,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":46997,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage16.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/f45acdd6b81b5e6651dfb66f.png"},{"id":93566338,"identity":"35859ca2-c956-4804-bafc-18459eb4098b","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"png","order_by":30,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":18514,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage17.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/af90ec72aceb606fecf2f38e.png"},{"id":93567168,"identity":"5393b720-8c15-421e-838c-1f7c25b48258","added_by":"auto","created_at":"2025-10-15 08:38:47","extension":"png","order_by":31,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":18626,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage18.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/49fe45646c24235a885cf0d6.png"},{"id":93567170,"identity":"1b844d0f-3d2a-47ae-b0d2-e1a1bd2b1d60","added_by":"auto","created_at":"2025-10-15 08:38:47","extension":"png","order_by":32,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":18189,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage19.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/fc8f6006693c1173fd9de556.png"},{"id":93567172,"identity":"4da50247-ba7c-4daa-a8f6-71c49d2cec1d","added_by":"auto","created_at":"2025-10-15 08:38:47","extension":"png","order_by":33,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":44642,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/696b9e1d8913d6033778d192.png"},{"id":93567161,"identity":"573fda33-4d5b-4857-86ba-df7b85614986","added_by":"auto","created_at":"2025-10-15 08:38:46","extension":"png","order_by":34,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":40849,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/95518d6b682fddea8c6b112c.png"},{"id":93566346,"identity":"63c417e1-8afe-4dbb-b469-d1a4d65072b1","added_by":"auto","created_at":"2025-10-15 08:30:47","extension":"png","order_by":35,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":40001,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/f226bf81467e22fc2022f50c.png"},{"id":93566347,"identity":"057bbcc0-4304-420e-9a77-9fd181a55b89","added_by":"auto","created_at":"2025-10-15 08:30:47","extension":"png","order_by":36,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":37036,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/ae2377a2cef5225defc5045e.png"},{"id":93566354,"identity":"4e325b84-9562-4225-8cc5-ffe9119c3ae7","added_by":"auto","created_at":"2025-10-15 08:30:47","extension":"png","order_by":37,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":40354,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/587d22f35eff7818a7017b9e.png"},{"id":93566336,"identity":"9dc16cd1-0aac-45c2-bd5e-968dc3b71171","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"png","order_by":38,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":46866,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/fa7087b9768eb04b1a7a75a7.png"},{"id":93566360,"identity":"e3df44ec-cb86-47b3-b8e2-f89590676d8b","added_by":"auto","created_at":"2025-10-15 08:30:47","extension":"png","order_by":39,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":42000,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/433bc681a00c0e4f7aefbe4f.png"},{"id":93566345,"identity":"b9a1e22f-cf30-4efb-abec-e0d0a1cccce9","added_by":"auto","created_at":"2025-10-15 08:30:47","extension":"png","order_by":40,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":44493,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/bb62e6373dca50c178a1e84e.png"},{"id":93566356,"identity":"4e91ca6a-f332-4e00-8732-219453c319e8","added_by":"auto","created_at":"2025-10-15 08:30:47","extension":"xml","order_by":41,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":210792,"visible":true,"origin":"","legend":"","description":"","filename":"9c11c1c35f2e4d9c97417515b1197f1d1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/e07d6592c7820d30d2ebcf75.xml"},{"id":93567166,"identity":"70972d06-c727-4cdd-981b-dadbd4089b9f","added_by":"auto","created_at":"2025-10-15 08:38:46","extension":"html","order_by":42,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":218162,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/fe2647e1fbb29b1d51a0ea57.html"},{"id":93566316,"identity":"1f019929-8627-4192-b4f1-1b36d058db5f","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":251426,"visible":true,"origin":"","legend":"\u003cp\u003eOverview of the methodological procedures\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAlt text: Flowchart illustrating the methodological procedures for analysing Instagram content from city tourism accounts such as @visitlondon, @parisjetaime, and others. The process includes extracting Instagram posts for captions, image URLs, post URLs, likes, comments, and followers, followed by image captioning using Open Clip Torch, mapping hedonic and utilitarian scales, and calculating engagement rates. Sentence transformers and BERTopics are used for dimensionality reduction, topic extraction, and projection into a scales subspace. The data undergoes normalisation, regression analysis, and multicollinearity removal using a negative binomial model.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/46ab5a96171af3a1d4993adf.png"},{"id":93566313,"identity":"f5615878-b997-471b-a177-ed58ae23991c","added_by":"auto","created_at":"2025-10-15 08:30:46","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":174425,"visible":true,"origin":"","legend":"\u003cp\u003eTopics listed under urban exploration experiences\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAlt text: Five-line graphs comparing positive and negative emotional attributes across different categories: Outdoor Gathering, Church and Architecture, Dining Ambience, Ferris Wheel and Festivals, and Castle Architecture. Each graph has a horizontal axis ranging from -0.2 to 0.2 and a vertical axis listing emotional attributes, with blue data points connected by red lines to visualise variations in emotional attributes for each category.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/127677414a16cf27ae96d7c6.png"},{"id":93566311,"identity":"a161ddc3-8d7c-4530-aa79-ebd704658005","added_by":"auto","created_at":"2025-10-15 08:30:45","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":177762,"visible":true,"origin":"","legend":"\u003cp\u003eTopics listed under scenery and leisure retreats\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAlt text: Five-line graphs comparing positive and negative emotional attributes across different categories: Art and sculpture, floral impressions, nature and parks, riverside charm, and cityscape and tranquil scenery. Each graph has a horizontal axis ranging from -0.2 to 0.2 and a vertical axis listing emotional attributes, with blue data points connected by red lines to visualise variations in emotional attributes for each category.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/fa100a8b3093e4bee98718ea.png"},{"id":93567156,"identity":"4021f027-1d24-4906-bd32-d459854a68a9","added_by":"auto","created_at":"2025-10-15 08:38:46","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":170778,"visible":true,"origin":"","legend":"\u003cp\u003eTopics listed under multimodal excursions\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAlt text: Five-line graphs comparing positive and negative emotional attributes across different categories: Boat tours, railway exploration, cycling adventures, elevated urban transit, and carriage horse rides. Each graph has a horizontal axis ranging from -0.2 to 0.2 and a vertical axis listing emotional attributes, with blue data points connected by red lines to visualise variations in emotional attributes for each category.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/ca4de3a7605444ecd16cfdd7.png"},{"id":108437709,"identity":"2d8bfb79-3f19-450c-9ade-2282a022f008","added_by":"auto","created_at":"2026-05-04 16:02:43","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1271079,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7591283/v1/0d5199b7-6894-4b23-893c-815fdc34abd0.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Feeling travel photographs: Emotional labelling in machine-generated semantic differentials","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eThe rise of social media has led to a growing emphasis on visualisation (Xu et al., \u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Particularly, photographs have become the backbone, with platforms such as Instagram relying heavily on visual content to attract and engage with users (Hauser et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Driven by the evolution of experience design, marketers now focus on creating visually engaging content that elicits emotional arousal in a hedonic context (Gon, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Among various types of visually centred multimedia, such as images, livestreaming, and videos (Fu et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), pictures play a crucial role (Yu \u0026amp; Egger, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). In the tourism context, pictures are key to understanding the emotional experiences conveyed through visual elements (Vu et al., \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Visual content not only showcases product features and service details, but also effectively conveys hedonic values that elicit emotional responses (Deng \u0026amp; Li, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). For instance, visual elements of tourism pictures such as colours and objects have specific associations that can evoke diverse feelings in viewers (L\u0026oacute;pez-Chao \u0026amp; Lopez-Pena, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Yu \u0026amp; Egger, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eIn fact, the discussion above aligns with the broader dichotomy of hedonic and utilitarian experiences in consumption emotions (Davari et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). In tourism, hedonic consumption refers to experiences that provide pleasure and emotional satisfaction (Soldat et al., \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), such as the enjoyment of beautiful scenery or cultural activities. Utilitarian consumption, on the other hand, focuses on practicality and functional benefits (Lee et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), like the convenience and efficiency of travel services. Balancing practical tourism information (Hong et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) with the pleasure-seeking aspect of social media and tourism (Virtanen et al., \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e2017\u003c/span\u003e) is salient. This balance affects how tourists engage with visual content and interact with digital media (Li \u0026amp; Xie, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Nonetheless, despite the increasing interest in understanding how photographs contribute to overall experiences (Picazo \u0026amp; Moreno-Gil, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), most research in this domain centres on the hedonic aspects within the realm of tourism marketing (Filieri et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Yu \u0026amp; Egger, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Although the experiential impact of visuals in tourism marketing has been recognised (Yu \u0026amp; Egger, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), the functional angles, often encapsulated in utilitarian value, remain underexplored. Furthermore, recent scholars have underpinned that emotional aspects within the tourism imagery are still not thoroughly explored, mainly due to the challenge of objectively measuring the intangible nature of emotions (Vu et al., \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eMethodologically speaking, previous studies on hedonic and utilitarian values primarily relied on semantic differential techniques with traditional survey methods (Kuo et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Lee et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Li et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). While useful for capturing opinions on bipolar scales, these surveys face criticisms like social desirability bias and central tendency errors (Br\u0026uuml;hlmann et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). More advanced approaches involve data analytics to examine how visual elements influence user engagement on social media (Rietveld et al., \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Despite advancements in visual data analytics, researchers often rely on previous literature for interpretation (Yu \u0026amp; Egger, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) and focus mainly on basic emotional responses (Vu et al., \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). This highlights the need for a new approach to better understand the experiences conveyed by visual elements (Mathew et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). One way to measure the underlying emotions in visual content is the use of machine learning to generate interpretable bipolar dimensions (Reelfs et al., \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). This cutting-edge technique in data analytics (Engler et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) shows promise for understanding experiential marketing of tourism destinations and informing effective strategies.\u003c/p\u003e\u003cp\u003eTherefore, this research aims to investigate the intertwining of hedonic and utilitarian aspects of tourism-related pictures shared by marketers and their relationship with user engagement using machine-generated semantic differentials. By bridging data analytics and visual content marketing, this research is novel in that it uses an interdisciplinary approach to examine the emotional connotations of pictures on user engagement in social media. Specifically, this study advances the literature by exploring the complex relationship between hedonic and utilitarian consumption in visual tourism marketing across various themes. Using machine-generated semantic differentials provides a new and more objective way to interpret emotional responses to images. Practically, these insights help marketers balance emotional appeal and craft effective visual strategies to boost engagement.\u003c/p\u003e"},{"header":"2. Literature review","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1. The effect of visual content on user engagement\u003c/h2\u003e\u003cp\u003eIn the broader context of multimedia marketing, text, videos, and images all play significant roles in engaging consumers (Wang \u0026amp; Luo, \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). However, pictures have emerged as particularly influential in visual marketing due to their immediacy and emotional impact (Yu \u0026amp; Egger, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). The use of various pictorial elements play a critical role in the creation of visual content, as they influence how consumers perceive the images (Xu et al., \u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). For instance, the colour scheme (Yu \u0026amp; Egger, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), message appeals (Rietveld et al., \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), and image quality (Li \u0026amp; Xie, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) can significantly impact the audience\u0026rsquo;s emotional response and engagement in digital communication. Building on the principles of experience design, marketers are now striving to intentionally create content and incorporate design thinking (Tussyadiah, \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2014\u003c/span\u003e) when developing marketing materials.\u003c/p\u003e\u003cp\u003eThrough carefully crafted visual content, marketers seek to create a more meaningful and authentic relationship with online users (Wies et al., \u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), thereby fostering emotional connections that can lead to increased engagement (Lim et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Particularly, Instagram has become an essential platform for tourism marketing (Hauser et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). The platform is known for its heavy reliance on visuals, with photos and videos being the primary means of communication. On Instagram, evaluation of marketing campaigns relies on user engagement, which reflects how well a post resonates with potential tourists (Yu \u0026amp; Egger, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). The observable metrics is typically shown by the number of likes and comments (Rietveld et al., \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Notably, differentiating likes and comments as dimensions of user engagement is crucial because they infer various level of interaction and interest that users have (Wies et al., \u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eInterestingly, recent research revealed that pictures featuring a special or unique activity tend to receive more likes, whereas content requiring a higher level of awareness (e.g., environmental issues) generates more comments (Aramendia-Muneta et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). In the context of sport events, embedding cultural elements in the pictures lead to higher engagement, while nature-related pictures yield the fewest likes and comments (Taberner \u0026amp; Junc\u0026agrave;, \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). However, these studies primarily focus on the surface-level objects within images (Taecharungroj \u0026amp; Mathayomchan, \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Yu et al., \u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), leaving the emotional dimensions embedded in these image objects across different contexts underexplored. Scholars have suggested that visual content can indeed convey a wide range of emotional experiences (Yu \u0026amp; Egger, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). For instance, photos of serene beach scenes may evoke peace and relaxation, while images of waves can generate excitement or even fear (Vu et al., \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Seeing that digital experiences increasingly focus on eliciting emotional responses, it is no longer sufficient for visual elements to be merely appealing (Hauser et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Instead, they need to (implicitly) convey emotional experiences that enable potential tourists to resonate and form a connection with the content (Xiao et al., \u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e2.2. Hedonic and utilitarian consumption in digital marketing\u003c/h2\u003e\u003cp\u003eThe evaluation of emotional dimensions in visual content aligns seamlessly with the framework of consumption values, encompassing hedonic and utilitarian experiences (Davari et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Hedonic experiences are experiential and enjoyable in nature (Soldat et al., \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), such as the feelings of excitement or relaxation that consumers derive from leisure activities or holiday (Su et al., \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). In contrast, utilitarian consumption centres around practical benefits and functionality, emphasising aspects such as usefulness and efficiency (Roggeveen et al., \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Utilitarian values also involve an appreciation for the practical benefits by assessing whether a product or service can aid in achieving specific goals (Lee et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eRevisiting foundational consumption theories highlights the critical role they play in evaluating how visuals offer marketers opportunities to showcase their products and services. This can be achieved through diverse presentation forms that leverage both hedonic and utilitarian approaches (Su et al., \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). While it appears that there has been a shift towards a more experiential and emotional emphasis in tourism, as consumers increasingly seek out hedonic value and deeper emotional attachments to enhance the overall experience (Tussyadiah, \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2014\u003c/span\u003e), it is important to acknowledge that many experiences concurrently serve functional purposes. Taking transportation as an illustrative example, transportation services, such as those offered by ride-sharing platforms or travel agencies, often highlight not only the joy of the journey but the utilitarian efficiency and convenience (Yuda Bakti et al., \u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eOn social media, even as companies tap into the hedonic aspect of experience (e.g., happiness, excitement) to craft successful marketing campaigns (Liadeli et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), functional elements in digital materials are not to be underestimated (Chen \u0026amp; Fu, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Certainly, digital content mainly serving functional purposes may be less effective in triggering engagement (Liadeli et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Yet, although visuals might emphasise the hedonic allure of an environment, it simultaneously addresses practical considerations (Chen \u0026amp; Fu, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) such as comfort, amenities, and accessibility. Hence, the evaluation of these emotional nuances becomes integral to comprehending how consumers extract value from their interactions with diverse forms of visual content. To do so, the emergence of visual analytics such as image mining (Ma \u0026amp; Palacios, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), clustering (Arefieva et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), and segmentation (Nguyen et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), has become increasingly important by providing an effective way that enables marketers to interpret visual data (Aramendia-Muneta et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e2.3. Machine-generated semantic differentials\u003c/h2\u003e\u003cp\u003eRooted in psychology, one way to interpret the data is the application of the semantic differential (Mindak, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e1961\u003c/span\u003e), which is used to measure and analyse one\u0026rsquo;s perceptions, attitudes, and emotions towards various stimuli such as products, services, and brands (Huang et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Conventionally, this technique involves asking individuals to rate the concepts being tested along a series of bipolar scales (Mathew et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), such as good-bad and strong-weak, which offers insights into the underlying attitudes and emotions that people associate with the stimuli (Huang et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). One of the key advantages of the semantic differential is that it provides a structured approach to measuring subjective attitudes and perceptions (Lin et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe semantic differential has been adopted extensively in marketing research to measure attitudes and preferences (Huang et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Lim \u0026amp; Childs, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). For example, researchers have used it to analyse consumers\u0026rsquo; perceptions of product design (Huang et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2012\u003c/span\u003e), to understand their attitudes towards brands (Lim \u0026amp; Childs, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), and to measure the effectiveness of social media advertisements (Sung et al., \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). One area where the semantic differential has been particularly useful is in understanding emotions (Das \u0026amp; Varshneya, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Huang et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2012\u003c/span\u003e) or hedonic and utilitarian value (Voss et al., \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2003\u003c/span\u003e). By using bipolar scales that capture affective terms such as dull-exciting, enjoyable-unenjoyable, it is possible to obtain a better understanding of the emotional responses a person may be undergoing (Das \u0026amp; Varshneya, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Voss et al., \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2003\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eGiven the swift accumulation of data, recent advancement in data analytics has shown promising results in interpreting pre-trained word embeddings (Mathew et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Echoing semantic differential, the learning model relies on the concept of polar opposites, where words are paired with their antonyms to enable interpretability of the embeddings (Reelfs et al., \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Its application can be seen in research focusing on sentiment analysis (Engler et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). For example, Reelfs et al. (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) transformed arbitrary word embeddings into interpretable counterparts for emotion detection based on visual symbols (e.g., emojis). However, one limitation is that the embedding information in the digital landscape (Mathew et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) may not fully capture the complexity of consumer behaviour in specific contexts. Moreover, the use of pre-trained word embeddings may not accurately reflect the unique nuances in meaning that are specific to different industries, products, and services (Mathew et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Therefore, it is essential to conduct research that incorporates how different marketing context may influence consumer behaviour (Stremersch et al., \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e\u003c/div\u003e"},{"header":"3. Methodology","content":"\u003cp\u003eThis study employs a systematic approach to explore the emotional and experiential dimensions of tourism-related imagery on social media and their impact on user engagement. The research is structured into five main steps: 1) destination selection, 2) data extraction, 3) image captioning and topic modelling, 4) machine-generated semantic differentials, and 5) analysis of user engagement. The detailed procedures follow. An overview of the methodological procedures is presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e3.1. Step 1: Destination selection\u003c/h2\u003e\u003cp\u003eThis study gathered data from the most popular destinations in different European countries. Specifically, destinations were selected based on the top 15 cities having the highest number of total bed-nights reported in the 18th edition of the CityDNA Benchmarking Report in Europe (City Destinations Alliance, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Yet, to maximise the diversity of destination characteristics, the second-ranked city in the same country was excluded. Cities without an official Instagram account from the destination marketing organisation were also excluded from further analysis. This led to a total of ten destinations in ten different countries, as summarised in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eSummary of the selected destinations (\u003cem\u003eas of March 2023\u003c/em\u003e)\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCity\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eBednights\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAccount\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003ePost (n)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eFollower (n)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLondon\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e25,542,000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e@visitlondon\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e3,981\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1.6M\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eParis\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e20,468,000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e@parisjetaime\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e3,490\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e687K\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBerlin\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e13,983,000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e@visitberlin\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e2,419\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e444K\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMadrid\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e10,932,000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e@visita_madrid\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e1,902\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e379K\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eStockholm\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e9,233,000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e@visitstockholm\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e4,680\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e345K\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAmsterdam\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e5,776,000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e@iamsterdam\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e4,464\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e322K\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eVienna\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e5,407,000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e@vienna\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e3,588\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e501K\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePrague\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e5,257,000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e@cityofprague\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e603\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e115K\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMilan\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e5,256,000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e@visit_milano\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e2,560\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e168K\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLisbon\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e5,186,000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e@visit_lisboa\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e4,691\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e133K\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003e3.2. Step 2: Data extraction\u003c/h2\u003e\u003cp\u003eData extraction was conducted in March 2023 using Apify, a web scaping platform, for all available posts of each account on Instagram, leading to a total of 35,459 posts. The collected data included post captions, date of the post, type of post (pictures/videos), number of likes and comments, post URLs, and image URLs. After removing 2,346 video-based posts, all pictures were downloaded based on the image URLs. Yet, since some of the pictures were duplicated and around 100 images were unavailable for download, potentially because they were deleted by the respective users, those records were excluded. The final dataset consisted of 33,001 picture-based posts.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003e3.3. Step 3: Image captioning and topic modelling\u003c/h2\u003e\u003cp\u003eBased on the extracted pictures, the next step involves classifying visual content based on their characteristics to gain deeper insights into different tourism experiences. First, image captioning, a technique for object recognition, was conducted using Open Clip Torch (Radford et al., \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), an open-source method from OpenAI. Unlike traditional image annotations that provide simple labels, this method generates short, descriptive captions for each image, such as \u0026ldquo;\u003cem\u003etwo people cross a street in front of a building\u003c/em\u003e\u0026rdquo;, offering a connected narrative rather than a list of disconnected labels like \u0026ldquo;\u003cem\u003epeople\u003c/em\u003e,\u0026rdquo; \u0026ldquo;\u003cem\u003ebuilding\u003c/em\u003e,\u0026rdquo; and \u0026ldquo;\u003cem\u003estreet\u003c/em\u003e.\u0026rdquo; This approach enables an automatic understanding of scene context, location, objects, people, and their interactions (Hossain, 2019). Moreover, the syntactically coherent textual descriptions produced with this tool integrate seamlessly with natural language processing techniques (You, 2016) that rely on context-sensitive embeddings for subsequent analysis.\u003c/p\u003e\u003cp\u003eThereafter, BERTopics (Grootendorst, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) were used to extract themes from the automatically generated image captions. This process employed sentence transformers (Reimers \u0026amp; Gurevych, \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), to create context-sensitive embeddings using a pretrained large language model (LLM) (i.e., the \u0026ldquo;all-MiniLM-L6-v2\u0026rdquo; model) due to its strong performance in natural language processing tasks like sentiment analysis, summarisation, named entity recognition, and translation (Liu, 2020). Applying these transformers to the image captions embedded each image into a 384-dimensional vector space. In this space, images with similar captions tend to be located close to each other, indicated by small angles between their vectors. These caption vectors were then utilised for topic modelling to group images into thematically similar categories and to relate them to the hedonic and utilitarian aspects of visual content (see Section 3.4).\u003c/p\u003e\u003cp\u003eSubsequently, Uniform Manifold Approximation and Projection (UMAP) were applied to reduce the dimensionality of sentence embeddings (McInnes et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). These lower-dimensional representations were then used for clustering the images. Only after the clustering process was completed did the pre-processing to refine topic representations occur. This involved removing stopwords, followed by applying the k-means algorithm to group similar vectors into clusters representing topics. Next, the term frequency-inverse document frequency (TF-IDF) scores, used to evaluate word relevance, were downscaled using BERTopics, which involves taking the square root of the scores to lessen the impact of common words. Each word within a topic was assigned a class-based TF-IDF score, with the top words being those with the highest scores. Ultimately, 15 distinct topics were extracted, showing clear separation between clusters. To assess the clustering quality, the silhouette coefficient was calculated, yielding a value of 0.36, indicating a moderate level of clustering effectiveness.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\u003ch2\u003e3.4. Step 4: Image captioning in machine-generated semantic differentials\u003c/h2\u003e\u003cp\u003eAfterwards, this study applied semantic differential as the measurement scale, and the bipolar dimensions were visualised in a framework for each of the identified topics. The bipolar measurement is based on the hedonic/utilitarian consumption scale (Voss et al., \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2003\u003c/span\u003e). The hedonic elements include fun\u0026ndash;not fun, exciting\u0026ndash;dull, delightful\u0026ndash;not delightful, thrilling\u0026ndash;not thrilling, enjoyable\u0026ndash;unenjoyable, happy\u0026ndash;not happy, pleasant\u0026ndash;unpleasant, playful\u0026ndash;not playful, cheerful\u0026ndash;not cheerful, amusing\u0026ndash;not amusing, sensuous\u0026ndash;not sensuous, and funny\u0026ndash;not funny. The utilitarian elements are effective\u0026ndash;ineffective, helpful\u0026ndash;unhelpful, functional\u0026ndash;not functional, necessary/unnecessary, practical\u0026ndash;impractical, beneficial\u0026ndash;harmful, useful\u0026ndash;useless, sensible\u0026ndash;not sensible, efficient\u0026ndash;inefficient, unproductive\u0026ndash;productive, handy\u0026ndash;not handy, and problem solving\u0026ndash;not problem solving.\u003c/p\u003e\u003cp\u003eFor the positive poles, the researchers extracted their definitions from the WordNet (Fellbaum, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2005\u003c/span\u003e), a comprehensive lexical database for English. WordNet definitions are typically brief sentences; for example, \u0026ldquo;fun\u0026rdquo; is defined as \u0026ldquo;activities that are enjoyable or amusing,\u0026rdquo; and \u0026ldquo;cheerful\u0026rdquo; as \u0026ldquo;being full of or promoting cheer; having or showing good spirits.\u0026rdquo; These definitions were transformed using the same sentence transformers applied to the automatically extracted picture captions, embedding the semantic differentials as dense vector representations within a shared 384-dimensional space. Each positive pole is embedded in this space, with negative poles represented as the negatives of their positive counterparts. For each scale, the pole vectors span a subspace within this 384-dimensional space, allowing the projection of image caption vectors onto these subspaces. Each scale comprising twelve poles, resulting in a 24-dimensional subspace. By projecting the picture captions onto these semantic differential subspaces, the study generated dense vector representations that link image captions with the hedonic and utilitarian scales, embedding the captions into the same vector space defined by these scales.\u003c/p\u003e\u003cp\u003eNotably, this study performed correlation analysis on the semantic differentials before the projection was calculated. The correlation analysis revealed weak to moderate correlations with stronger correlations within the hedonic (maximal correlation was 0.78 between \u0026ldquo;pleasant\u0026rdquo; and \u0026ldquo;enjoyable\u0026rdquo; and minimal correlation was 0.25 between \u0026ldquo;playful\u0026rdquo; and \u0026ldquo;enjoyable\u0026rdquo;) than within the utilitarian subscale (maximal correlation was 0.63 between \u0026ldquo;productive\u0026rdquo; and \u0026ldquo;effective\u0026rdquo; and minimal correlation was 0.09 between \u0026ldquo;beneficial\u0026rdquo; and \u0026ldquo;problem solving\u0026rdquo;). The correlations between the subcomponents of the different subscales were substantially weaker. Although some of the correlations can be considered strong, this study eliminated these collinearities within the subsequent regression analysis.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003e3.5. Step 5: Analysis of user engagement across different marketing contexts\u003c/h2\u003e\u003cp\u003eAfter preparing the pre-trained word vectors, the analysis proceeded to assess user engagement through multiple regression analysis. Since likes and comments reflect different levels of engagement, they were examined separately (Aramendia-Muneta et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). The regression analysis began by eliminating collinearities in hedonic and utilitarian subscales using the variance inflation factor (VIF), removing components with a mean VIF over four (Wu \u0026amp; You, \u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Since likes and comments are count data, a negative binomial regression model was employed. To normalise engagement metrics, likes and comments were divided by the number of followers. In the model, this normalisation is achieved by setting an offset (exposure) parameter, which adjusts for differences in follower counts. The model then automatically accounts for this by dividing the outcome variables by the number of followers. Because the engagement data were not normally distributed, a logarithmic transformation was applied. Thus, the final regression models were multiplicative, with coefficients interpreted as relative percentage differences from baseline levels. Notably, negative binomial models assume equal mean and variance, but overdispersion often occurs when variance exceeds the mean. To address this, the overdispersion parameter was adjusted to scale the data appropriately. The study determined the optimal settings by fitting models across a range of overdispersion values and selecting the one closest to one. Lastly, the goodness-of-fit of all models was evaluated using chi-squared tests on the likelihood ratio comparing each model to the baseline.\u003c/p\u003e\u003cp\u003eWhile all available data has been collected, early years showed mostly zero comments and likes. To mitigate potential bias, the subsequent analysis focused on the years from 2016 to 2022. Due to seasonality in tourism, the data were split into summer and winter seasons (summer\u0026thinsp;=\u0026thinsp;0, winter\u0026thinsp;=\u0026thinsp;1) to account for potential variations in user preferences. The seasonality in demand also shows impact on online engagement (Villamediana et al., \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). In this research, summer spans from April through September and winter is between October and March. Next, four regression models were run altogether: for both engagement types (i.e., likes and comments) and with all 15 topics as control variables and without topics. Multicollinearity analysis included six elements from the hedonic/utilitarian scales: three hedonic (fun, thrilling, sensuous) and three utilitarian (functional, useful, productive) (see Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e\u003c/div\u003e"},{"header":"4. Results","content":"\u003cp\u003eThe findings are structured into four distinct sections. The first presents topic modelling results from image captions to overview tourism-related picture attributes. The second uses semantic differentials to visually represent bipolar dimensions of each topic, highlighting hedonic and utilitarian aspects. The final two sections explore the interaction effects of these facets on user engagement.\u003c/p\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003e4.1. Topic modelling based on image captions\u003c/h2\u003e\u003cp\u003eBased on the BERTopics, this study reveals 15 clusters featuring different pictorial contexts. Topics were named using TF-IDF keywords and representative pictures. Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e shows the clusters, organised into three groups based on the inter-topic distance map: urban exploration experiences (featuring urban elements like churches and Ferris wheels), scenery and leisure retreats (focused on nature, art, and atmosphere), and multimodal excursions (involving transportation modes like cycling and railways).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eSummary of the BERTopics results\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCluster\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eTopic\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eKeywords\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003en\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"4\" nameend=\"c4\" namest=\"c1\"\u003e\u003cp\u003e\u003cb\u003eGroup 1: Urban Exploration Experiences\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eOutdoor gathering\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eGroup, people, sitting, fountain, crowd, sidewalk, standing, walking, Eiffel, women\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e5,833\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDining ambience\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eTable, food, plate, wine, topped, store, cup, coffee, chairs, holding\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e2,706\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eChurch and Architecture\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eChurch, cathedral, car, cars, parked, clock, trolley, steeple, road, driving\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1,722\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eFerris wheel and festivals\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eChristmas, Ferris, wheel, tree, decorations, rain, lit, lights, day, night\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1,027\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e12\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCastle architecture\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eCastle, decker, double, windmill, bus, driving, clock, Brandenburger, tor, windmills\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e791\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"4\" nameend=\"c4\" namest=\"c1\"\u003e\u003cp\u003e\u003cb\u003eGroup 2: Scenery and Leisure Retreats\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eArt and sculpture\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eWoman, statue, man, person, mural, painting, standing, holding, dog, dress\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e4,878\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eFloral impressions\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eFlowers, sign, plants, room, door, windows, staircase, leading ceiling, archway\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3,946\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNature and parks\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eTrees, sky, park, rainbow, tree, background, garden, flag, pond, flags\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3,347\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eRiverside charm\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eSun, bridge, setting, bicycle, alley, way, parked, shining, river, bicycles\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1,977\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e14\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCityscape and tranquil scenery\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eSkyline, windows, tulips, sunset, square, houses, town, point, high, row\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e286\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"4\" nameend=\"c4\" namest=\"c1\"\u003e\u003cp\u003e\u003cb\u003eGroup 3: Multimodal Excursions\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eBoat tours\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eBoats, boat, docked, water, canal, body, harbor, pier, floating, aerial\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3,037\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eRailway exploration\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eTrain, tracks, station, travelling, tunnel, subway, track, platform, yellow, trains\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1,072\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCycling adventures\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eBike, riding, night, person, scooter, aerial, man, bicycle, street, colorful\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e904\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e11\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eElevated urban transit\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eTram, high, point, bus, vantage, going, decker, driving, double, dome\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e872\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCarriage horse rides\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eDrawn, carriage, horse, balconies, caf\u0026eacute;, horses, swan, swans, skyline, carriages\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e604\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003e4.2. Bipolar dimensions of pictorial attributes\u003c/h2\u003e\u003cp\u003eBuilding on identified topics, the study visualises their hedonic and utilitarian aspects using a novel framework for interpreting pre-trained embeddings. By averaging embeddings per topic, diagrams show how topics are positioned in polar coordinates based on hedonic and utilitarian scales (Voss et al., \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2003\u003c/span\u003e). These illustrate user reactions to specific image subjects, with each dimension representing opposing semantic pairs. Responses near the edge indicate a strong correlation, while those near the centre suggest a weaker connection.\u003c/p\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e consists of pictorial topics related to urban experiences. Interestingly, utilitarian aspects are prominent, as images of architecture and outdoor spaces evoke handiness and accessibility. When it comes to dining ambience, the images strive to convey sensory and delightful experiences. In Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, which depicts scenery and leisure topics, sensory consumption predominates, while purely amusing consumption is absent. An intriguing observation is that content producers appear to intentionally minimise practical feelings in nature-related topics. Turning to Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e, which encapsulates tourism experiences associated with commuting and transportation, the results unsurprisingly demonstrate an emphasis on conveying practical and problem-solving consumption. Moreover, there is a notable presence of sensuous consumption, highlighting the foundational role of sensory engagement for tourists.\u003c/p\u003e\u003cp\u003eHowever, a few unexpected observations have surfaced. For instance, the results suggest that pictures associated with Ferris wheels and festivals tend to be slightly less hedonic and convey more unpleasant consumption. Similarly, images featuring natural parks and flowers appear to connote unpleasant experiences. Given the prevailing strategies employed by destination marketers in presenting diverse tourism attributes, there is a need for further exploration into their effectiveness in influencing potential tourists\u0026rsquo; engagement with the posts, which will be addressed in the subsequent analysis.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003e4.3. Regression models: Initial exploration on hedonic/utilitarian scale\u003c/h2\u003e\u003cp\u003eThe regression models provide an overview of the contribution of various hedonic and utilitarian elements based on the number of likes and comments, respectively (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). First, the results suggest that seasonality did play a role in how users engage with the pictorial content. In both cases, there was a significant difference between winter (winter\u0026thinsp;=\u0026thinsp;1) and summer (winter\u0026thinsp;=\u0026thinsp;0), p\u003csub\u003elikes\u003c/sub\u003e\u0026lt;.001 and p\u003csub\u003ecomments\u003c/sub\u003e\u0026lt;.001, in which, pictures posted in winter usually received higher engagement rate potentially due to several holidays during the season in Europe such as Christmas and New Year.\u003c/p\u003e\u003cp\u003eOverall, the findings reveal that when pictures connoted a sense of thrill, there was a positive and significant difference in users\u0026rsquo; liking and commenting behaviour, p\u003csub\u003elikes\u003c/sub\u003e\u0026lt;.001 and p\u003csub\u003ecomments\u003c/sub\u003e=.001. Yet, pictures having functional positions had a significant and negative impact on the number of likes and comments, p\u003csub\u003elikes\u003c/sub\u003e\u0026lt;.001 and p\u003csub\u003ecomments\u003c/sub\u003e\u0026lt;.001. Interestingly, while embedding fun elements in pictures significantly encouraged one\u0026rsquo;s liking behaviour, p\u0026thinsp;\u0026lt;\u0026thinsp;.001, it was ineffective in triggering a deeper level of engagement as reflected by the number of comments, p\u0026thinsp;=\u0026thinsp;0.02. Furthermore, the results suggest that delivering a sense of usefulness, sensibility, and playfulness had a significant and negative effect on users\u0026rsquo; intention to comment on a post. Due to the diversity of pictorial content, the subsequent analysis delves deeper into the varied effects of hedonic and utilitarian elements in different settings.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eSummary of regression models without considering topics\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eLiking behaviour\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e\u003cp\u003eCommenting behaviour\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ecoef.\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003ez-score\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003ecoef.\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003ez-score\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eIntercept\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-4.3930***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-462.79\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-9.1170***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-737.078\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFunctional\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.0499***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-6.593\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-0.0503***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-5.107\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNecessary\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.0004\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.049\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-0.0053\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-0.554\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eUseful\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.0082\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-1.071\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-0.0431***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-4.344\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSensible\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.0092\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-1.208\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-0.0273**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-2.757\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eProductive\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.0745***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-9.322\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.0074\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.712\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFun\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.0351***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e4.590\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-0.0232*\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-2.328\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDelightful\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.0058\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-0.753\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.0127\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1.255\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eThrilling\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.0396***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e5 .088\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.0345**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e3.398\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePlayful\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.0147\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-1.882\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-0.0350**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-3.453\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCheerful\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.0177*\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-2.293\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-0.0270*\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-2.677\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSensuous\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.0183**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-2.571\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-0.0001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-0.012\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFunny\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.0079\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-1.038\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.0019\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.190\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSeason\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.1723***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e12.862\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.1550***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e8.887\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"5\"\u003e\u003cem\u003eNote\u003c/em\u003e: *p\u0026thinsp;\u0026lt;\u0026thinsp;.05; **p\u0026thinsp;\u0026lt;\u0026thinsp;.01; ***p\u0026thinsp;\u0026lt;\u0026thinsp;.001\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\u003ch2\u003e4.4. Regression models: Model selection and interactions\u003c/h2\u003e\u003cp\u003eAccording to the initial exploration of the scale (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e), this study then focuses on the covariates that were significant. Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e outlines the results of model selection on users\u0026rsquo; liking and commenting behaviours with a significant effect. As topics were used as control, the models explain how average predictions changed across topics when the bipolar coordinate was held constant. Notably, Topic 0 served as the baseline. Thus, when every bipolar dimension was 0 and the Topic was 0, the intercept inferred the average engagement. The coefficients for all other topics were always relative to this baseline. That is, if they were positive, they were associated with an increase in the average engagement as compared to Topic 0, and vice versa.\u003c/p\u003e\u003cp\u003eThe findings indicate that most of the topics significantly influenced the number of likes in a positive way, except Topic 1 (Art and sculpture), 5 (Dining ambience), and 13 (Carriage horse rides). As for commenting behaviour, significant predictions were observed mostly in topics related to transportation [Topic 4 (Boat tours), 8 (Railway exploration), and 11 (Elevated urban transit)], in addition to Topic 5 (Dining ambience) and 7 (Church and architecture).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eModel selection for liking and commenting behaviour\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"7\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eLiking behaviour\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"4\" nameend=\"c7\" namest=\"c4\"\u003e\u003cp\u003eCommenting behaviour\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ecoef.\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003ez-score\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003ecoef.\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003ez-score\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eIntercept\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-4.4580***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-248.136\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eIntercept\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e-9.1615***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e-374.13\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.0124\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-0.498\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.0010\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.030\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.0765**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-2.882\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e-0.0670\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e-1.860\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.0820**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e3.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.0213\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.574\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.1819***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e6.727\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.0758*\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e2.051\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.0050\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e0.167\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.2866***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e7.069\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.0810**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e2.555\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e-0.0061\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e-0.141\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.1668***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e5.104\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.1240**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e2.784\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.3097***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e7.963\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.1469**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e2.783\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.2240***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e5.546\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.0981\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e1.787\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.1120**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e2.723\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.0117\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.209\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 11\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.2563***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e6.058\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 11\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.1594**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e2.770\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 12\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.1716***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e3.898\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 12\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.0645\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e1.076\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.0636\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e1.306\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.0464\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.699\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 14\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.2130**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e3.083\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 14\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.1307\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e1.396\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"7\"\u003e\u003cem\u003eNote\u003c/em\u003e: *p\u0026thinsp;\u0026lt;\u0026thinsp;.05; **p\u0026thinsp;\u0026lt;\u0026thinsp;.01; ***p\u0026thinsp;\u0026lt;\u0026thinsp;.001\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eFurther exploration involved fitting the models with topic interactions. While the majority of interactions did not exhibit significant differences, the noteworthy insights emerged from those interactions that did prove to be significant (Table\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e), for functionality, usefulness, productiveness, thrill, and fun. First, in terms of the extent of functionality presenting in pictures, both Topic 5 (Dining ambience) and 10 (Cycling adventures) showed a negative impact on the number of likes and comments. Yet, there was a significant and positive relationship in one\u0026rsquo;s probability to like artistic pictures (Topic 1).\u003c/p\u003e\u003cp\u003eRegarding pictures embedded thrilling experiences, markedly, the number of likes and comments increased alongside the level of thrill in Topic 10 (cycling adventures). Conversely, the results demonstrate significant but negative relations with both liking and commenting behaviour in pictures featuring riverside charm (Topic 6) and railway exploration (Topic 8). Likewise, when dining ambience (Topic 5) contained a higher extent of thrill, it created adverse effect on users\u0026rsquo; intention to comment.\u003c/p\u003e\u003cp\u003eTurning to the dimension of usefulness, interestingly, significant and negative impacts were found in all cases. In general, pictures featuring Ferris wheel and festivals (Topic 9) received less likes and comments in relation to the extent of usefulness. Similar patterns were observed in Topic 1 (Art and sculpture), 2 (Floral impressions) and 5 (Dining ambience) on commenting behaviour, and Topic 10 (Cycling adventures) on liking behaviour.\u003c/p\u003e\u003cp\u003eConcerning the level of productivity, this study uncovers that connoting pictures with the quality of being productive encouraged one\u0026rsquo;s intention to like and comment on Topic 14 (Cityscape and tranquil scenery). This strategy was also effective in increasing the number of likes in pictures featuring floral impressions (Topic 2), railway exploration (Topic 8), and carriage horse rides (Topic 13). As for commenting behaviour, while embedding productive consumption in the pictures was effective for Topic 5 (Dining ambience), it decreased behavioural intention in Topic 7 (Church and architecture).\u003c/p\u003e\u003cp\u003eMoreover, to the extent that pictures connotated fun consumption, the models suggest some unexpected but thought-provoking observations. Particularly, for all topics listed under multimodal excursions (Topic 4, 8, 10, 11, and 13), the degree of fun correlated negatively with the number of likes. Similar patterns were also found in Topic 5 (Dining ambience), 9 (Ferris wheel and festivals), 12 (Castle architecture), and 3 (Nature and parks). Finally, when sensibility was analysed with the topic, no significant interaction effect was discovered.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eTopic interactions with significant effect\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"7\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eLiking behaviour*Topic\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"4\" nameend=\"c7\" namest=\"c4\"\u003e\u003cp\u003eCommenting behaviour*Topic\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ecoef.\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003ez-score\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003ecoef.\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003ez-score\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eFunctional\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u003cb\u003e-0.0216\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e\u003cb\u003e-1.247\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003eFunctional\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e-0.0388\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e-1.631\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.0875**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-2.881\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.0853**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e2.330\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.1655**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-3.303\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e-0.1006**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e-2.429\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e-0.1659**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e-2.441\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eUseful\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u003cb\u003e-0.0008\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e\u003cb\u003e-0.049\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003eUseful\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e0.006.\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e0.268\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.0887*\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-2.010\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e-0.0992**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e-3.032\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.1527**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-2.712\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e-0.0844*\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e-2.251\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e-0.0764*\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e-2.114\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e-0.1234*\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e-2.155\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eProductive\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u003cb\u003e-0.0943***\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e\u003cb\u003e-5.526\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003eProductive\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e-0.0206\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e-0.891\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.0637*\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e2.317\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.0792*\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e2.242\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.1298*\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e2.588\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e-0.1722**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e-2.667\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.1475*\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e2.408\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 14\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.5254**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e2.978\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 14\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.4926***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e4.064\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eThrilling\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u003cb\u003e0.0433**\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e\u003cb\u003e2.598\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003eThrilling\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e0.0716**\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e3.183\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.0989**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-3.098\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e-0.1680***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e-3.767\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.1996***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-4.22\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e-0.0992*\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e-2.355\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.1505**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e2.731\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e-0.1672**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e-2.643\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eTopic 10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.1391*\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e2.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eFun\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u003cb\u003e0.1062***\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e\u003cb\u003e7.542\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003eSensible\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e-0.0124\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e-0.577\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.0692**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-2.708\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.0803**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-2.622\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.0772**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-3.036\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.1912***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-4.521\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.1954***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-5.535\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.1287*\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-2.414\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 11\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.2547***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-3.912\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 12\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.1363*\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-2.202\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic 13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.1116*\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003e-2.187\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"7\"\u003e\u003cem\u003eNote\u003c/em\u003e: *p\u0026thinsp;\u0026lt;\u0026thinsp;.05; **p\u0026thinsp;\u0026lt;\u0026thinsp;.01; ***p\u0026thinsp;\u0026lt;\u0026thinsp;.001\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e"},{"header":"5. Discussion","content":"\u003cp\u003eIn a landscape where tourism experiences are conveyed visually (Xu et al., 2024), this research seeks to delve into the intricate relationship between the embedded hedonic and utilitarian aspects in images and their influence on user engagement. Nevertheless, it is important to recognise that this focus does not diminish the significance of other elements, such as post captions, popular hashtags, or alternative formats like videos. Notably, although this study centres on pictures, other additional components also play crucial roles in enhancing visibility, providing context, and boosting engagement (Fu et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Yu et al., \u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eInterestingly, unlike the commonly identified destination image attributes such as nature and outdoor environment (e.g., beach, mountain, forest, lake), gastronomy, architectures (e.g., museum, heritage, temple), as well as cityscape summarised in existing literature (Picazo \u0026amp; Moreno-Gil, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Yu \u0026amp; Egger, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), insights from image captions reveal a new facet on multimodal excursions. This includes boat rides, railway journeys, cycling, urban transit, and horse rides. These often-overlooked aspects contribute to a more comprehensive understanding of tourist experiences. However, in fact, the emphasis on the functional dimensions of tourism products, such as transportation modes, has been strengthened in research that analyses destination image through text mining based on online reviews (Li et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). This trend shows that traditional image attributes are extending to pictorial representations, highlighting that functional aspects, such as various transportation modes, are integral to overall travel experiences (K\u0026ouml;ltringer \u0026amp; Dickinger, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2015\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eSubsequently, the influence of different topics on potential tourists\u0026rsquo; liking and commenting behaviours is scrutinised. Concerning experiences related to transportation, these findings challenge preconceived notions regarding the universal appeal of fun in driving engagement (Vries et al., \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Despite tourism is about selling an emotional journey (Yu \u0026amp; Egger, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), this research initiates a reassessment of assumptions regarding user preferences and the impact of fun on shaping engagement across varied thematic contexts. Furthermore, although previous research found that incorporating activity-centric pictures can enhance user engagement (Aramendia-Muneta et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), it is important to consider the context. For multimodal excursions, users may unconsciously prefer informative content over entertaining appeals (Hong et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eWhile fun consumption was not prominent in multimodal excursions, thrill significantly increased user engagement in pictures of cycling adventures (Fossgard \u0026amp; Fredman, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). This heightened engagement likely stems from the dynamic nature of cycling, aligning with thrill-seeking expectations in adventure tourism (Kiatkawsin et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). This is further supported when considering functionality. A functional atmosphere in pictures might detract from desired experiences, discouraging engagement. Similarly, users might avoid engaging with serene visuals (e.g., riverside) if thrill elements are present, as they may prioritise tranquillity over excitement in nature-based tourism (Conti \u0026amp; Lexhagen, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Furthermore, the negative impact on commenting behaviour with dining experiences adds complexity to thrill consumption. This could be that tourists often seek relaxed, immersive gastronomic experiences (Dixit \u0026amp; Prayag, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), and introducing thrilling elements might disrupt this atmosphere, discouraging active commenting.\u003c/p\u003e\u003cp\u003eRegarding art and sculpture, the findings highlight the intricate interplay between functionality and thematic context. The perceived utility of art and sculpture can enhance user engagement rather than hinder it. When considering the multifaceted nature of art appreciation, functionality might be perceived as a means to deepen one\u0026rsquo;s understanding with the artistic elements (Liu, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). This integration acts as a bridge between aesthetic and practical dimensions (Botti, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2000\u003c/span\u003e), catering to diverse user preferences that find such content enriching and intellectually stimulating. Conversely, since art is often regarded as a form of expression and creativity (Botti, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2000\u003c/span\u003e), users may be less motivated to engage with content perceived as merely useful (Christiaans, 2002). These findings challenge the assumption that perceived usefulness always positively impacts user preferences (Aramendia-Muneta et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). In tourism marketing, which heavily relies on visual appeals (Arabadzhyan et al., 2021), user engagement dynamics may deviate from utility-based expectations (Chen \u0026amp; Fu, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). This aligns with the focus on experiential marketing, where tourists seek dynamic encounters that go beyond traditional notions of usefulness (Tussyadiah, \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eLastly, cityscape and tranquil scenery images highlight the productivity of visual content, prompting potential tourists to show appreciation through likes and comments. A potential explanation for this phenomenon is that individuals perceive a sense of productivity, akin to accomplishment (Leit\u0026atilde;o et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), when engaging with tranquil content. Such engagement, therefore, fosters a positive mental state, leading users to express their appreciation (Neuhofer et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). In another scenario, this study implies that users value content showcasing culinary productivity within dining ambience, such as the artistic and skilful aspects. This aligns with the prevailing trend of culinary exploration and visually appealing food experiences on social media (Gambetti \u0026amp; Han, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Nonetheless, users exploring church and architecture content might prioritise cultural significance and aesthetic qualities of the depicted locations (L\u0026oacute;pez-Chao \u0026amp; Lopez-Pena, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). In these contemplative settings, productive consumption may seem less relevant or disruptive to the experience.\u003c/p\u003e"},{"header":"6. Conclusion","content":"\u003cp\u003eEssentially, this research investigates the relationship between the hedonic and utilitarian aspects of tourism-related images shared on social media and their impact on user engagement. By applying a machine-generated semantic differential approach, it underscores that user engagement in tourism imagery is influenced by thematic context, challenging the assumption that fun universally drives engagement. For instance, functionality can enhance engagement with art and sculpture by deepening understanding, whereas perceived usefulness does not always correlate with engagement. Meanwhile, although thrill enhances engagement in cycling adventures, users may prefer informative content over entertaining appeals in multimodal excursions. Overall, the results provide a nuanced understanding of the interplay between emotional connotations in visual content and potential tourists\u0026rsquo; engagement on social media. The following sections delve deeper into the theoretical and practical implications.\u003c/p\u003e\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e\u003ch2\u003e6.1. Theoretical contributions\u003c/h2\u003e\u003cp\u003eThe interdisciplinary nature of this study makes theoretical advancements by bridging consumer psychology, data analytics, and emotional experience analysis, within the context of tourism visual marketing. Specifically, one of the key contributions lies in the introduction of a novel methodological approach that applies machine-generated semantic differentials to evaluate the interplay between hedonic and utilitarian values embedded in visual content. Given the inherent semantic variations in images, the application of machine learning in this study provides unique insights for future research. Distinct from traditional survey-based methods, one key advantage of the bipolar framework is its objectivity in measuring experiences (Br\u0026uuml;hlmann et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). By elevating the subjective aspects of semantic differentials to an objective level, the machine-generated approach enhances the robustness of the results.\u003c/p\u003e\u003cp\u003eBy transcending traditional approaches, this technique addresses the gap left by the lack of focus on emotional experiences in the analysis of tourism pictures (Vu et al., \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), highlighting the need for more comprehensive exploration. For instance, the dimension of usefulness challenges traditional assumptions, revealing that tourists in entertainment-oriented themes prioritise emotional and experiential content over practical information. The results also stimulate discussions regarding the broad popularity of enjoyment across different settings (Vries et al., \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e2012\u003c/span\u003e), prompting a re-evaluation of user preferences within different tourism contexts. Moreover, the use of data-driven techniques (Mathew et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) extends the status quo of the hedonic-utilitarian measurement (Voss et al., \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2003\u003c/span\u003e). By shedding light on the importance of hedonic and utilitarian consumption in social media marketing, this research emphasises the need to create visually appealing content that evokes various feelings, depending on the context, to attract and engage potential tourists (Gambetti \u0026amp; Han, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eAdditionally, beyond conventional themes commonly explored in previous literature, such as nature, dining, architecture, and cityscape (Picazo \u0026amp; Moreno-Gil, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), the identification of multimodal excursions (e.g., boat, railway, and horse rides) adds a new dimension to the visual representation of tourism experiences on social media. The utilisation of image captions for labelling also introduces an emerging approach in visual analysis research (He et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) as it potentially reveals underexplored topics in destination pictures. This finding challenges the existing paradigms that predominantly focus on traditional attributes such as nature and gastronomy (Conti \u0026amp; Lexhagen, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Picazo \u0026amp; Moreno-Gil, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Yu \u0026amp; Egger, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). By highlighting the emotional and experiential dimensions associated with these often-overlooked themes, this study encourages future research to explore the complexities of tourism experiences beyond conventional boundaries. Moreover, the findings encourage a re-evaluation of the hedonic-utilitarian dichotomy in consumption literature. They prompt scholars to explore the contextual factors influencing behaviour, enriching the theoretical discourse on emotional consumption in tourism and digital marketing. Overall, this research paves the way for the integration of new techniques beyond tourism, into other services and marketing disciplines, emphasising the importance of understanding and catering to diverse consumer preferences within specific thematic contexts for more effective and resonant strategies.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec20\" class=\"Section2\"\u003e\u003ch2\u003e6.2. Practical implications\u003c/h2\u003e\u003cp\u003eBy establishing a vital connection between visual elements and the efficacy of embedding diverse consumption experiences, this study offers destination marketers valuable insights for optimising social media content and engaging tourists effectively. For instance, one of the findings indicates that content centred around multimodal excursions tends to prioritise informativeness over entertainment. That is, marketers should focus on crafting visually engaging posts that emphasise the practical aspects of transportation experiences\u0026mdash;such as safety, efficiency, and accessibility\u0026mdash;while still conveying emotional resonance. In the realm of dining experiences, the study reveals a relationship between functionality and engagement. The results imply that marketers should consider integrating aspects of artistry and skill into their culinary visuals, as these attributes can foster deeper emotional connections. For example, posts that showcase chefs in action or beautifully plated dishes can enhance user appreciation and interaction.\u003c/p\u003e\u003cp\u003eThe study also emphasises the positive response to thrill in adventurous activities (e.g., cycling), indicating alignment with expectations in adventure tourism. Hence, content creators can leverage this insight by showcasing dynamic imagery that captures the excitement of cycling or boat tours. For instance, an Instagram post featuring a cyclist navigating through a vibrant landscape, paired with exhilarating captions, can effectively resonate with thrill-seekers and encourage higher interaction rates. However, marketers should exercise caution to avoid disrupting desired atmospheres, especially in serene environments. Moreover, exploring the merging of functionality with artistic expression provides another avenue for marketers. This trend reflects a preference for enhanced aesthetic experiences and interactive engagement, which is in line seamlessly with the contemporary emphasis on experiential marketing. In a nutshell, the insights from this research emphasise the necessity for marketers to adopt a tailored approach to their visual content strategies, recognising that the impact of hedonic and utilitarian elements varies significantly across different tourism contexts. Recognising these intricacies is crucial for destination marketers seeking to enhance user engagement and the overall effectiveness in visual marketing.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec21\" class=\"Section2\"\u003e\u003ch2\u003e6.3. Limitations and recommendations\u003c/h2\u003e\u003cp\u003eWhile this study provides valuable insights, it is not without its limitations. Firstly, despite the novelty of the data-driven bipolar framework, the study primarily focuses on the hedonic and utilitarian dimensions in isolation, potentially overlooking the full richness of tourism experiences. Future research is recommended to explore alternative measurement methods in experiential marketing in order to broaden the understanding of the impact of visual content on tourism. Researchers may also investigate the interactive effects and potential conflicts between these dimensions for a more comprehensive view. Furthermore, although efforts were made to ensure data source diversity, the focus on European destinations may fall short of capturing the unique cultures or atmospheres found in other regions. Additionally, since this study is conducted from the perspective of destination marketers, it is important to recognise that their viewpoint might differ from that of ordinary tourists. To address this limitation, future scholars should consider expanding the scope to include a more diverse range of destinations across various continents. Comparing these findings with tourists\u0026rsquo; perceived images could provide further insights. Meanwhile, in addition to analysing likes and comments, scholars are encouraged to delve deeper by examining the content of comments. This can be achieved through sentiment or emotional analysis to quantify user reactions, or by employing qualitative methods to gain richer insights. Finally, with the rise of short videos in the tourism domain, future studies are encouraged to delve into the distinctive attributes and effects of video content on user engagement. Additionally, factors such as post captions should be examined in tandem. This exploration would contribute to a more comprehensive understanding of the evolving landscape of visual content in marketing.\u003c/p\u003e\u003c/div\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eJoanne Yu: Conceptualization, Writing - Original Draft, Writing - Review \u0026amp; Editing; Denis Helic: Methodology, Analysis, Writing - Review \u0026amp; Editing; Astrid Dickinger: Conceptualization, Writing - Review \u0026amp; Editing; Markus Strohmaier: Methodology, Analysis\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAramendia-Muneta ME, Olarte-Pascual C, Ollo-L\u0026oacute;pez A (2020) Key Image Attributes to Elicit Likes and Comments on Instagram. J Promotion Manage 1\u0026ndash;27. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/10496491.2020.1809594\u003c/span\u003e\u003cspan address=\"10.1080/10496491.2020.1809594\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eArefieva V, Egger R, Yu J (2021) A machine learning approach to cluster destination image on Instagram. Tour Manag 85:104318. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.tourman.2021.104318\u003c/span\u003e\u003cspan address=\"10.1016/j.tourman.2021.104318\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBotti S (2000) What Role for Marketing in the Arts? An Analysis of Arts Consumption and Artistic Value. Int J Arts Manage, 14\u0026ndash;27\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBr\u0026uuml;hlmann F, Petralito S, Rieser DC, Aeschbach LF, Opwis K (2020) TrustDiff: Development and Validation of a Semantic Differential for User Trust on the Web. J Usability Stud 16(1):29\u0026ndash;48\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen JH, Fu J-R (2018) On the effects of perceived value in the mobile moment. Electron Commer Res Appl 27:118\u0026ndash;128. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.elerap.2017.12.009\u003c/span\u003e\u003cspan address=\"10.1016/j.elerap.2017.12.009\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCity Destinations Alliance (2022) \u003cem\u003eCityDNA Benchmarking Report\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://citydestinationsalliance.eu/citydna-benchmarking-report/\u003c/span\u003e\u003cspan address=\"https://citydestinationsalliance.eu/citydna-benchmarking-report/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eConti E, Lexhagen M (2020) Instagramming nature-based tourism experiences: a netnographic study of online photography and value creation. Tourism Manage Perspect 34:100650. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.tmp.2020.100650\u003c/span\u003e\u003cspan address=\"10.1016/j.tmp.2020.100650\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDas G, Varshneya G (2017) Consumer emotions: Determinants and outcomes in a shopping mall. J Retailing Consumer Serv 38:177\u0026ndash;185. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jretconser.2017.06.008\u003c/span\u003e\u003cspan address=\"10.1016/j.jretconser.2017.06.008\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDavari D, Nosrati S, Kim S (2024) Do cultural and individual values influence sustainable tourism and pro-environmental behavior? Focusing on Chinese millennials. J Travel Tourism Mark 41(4):559\u0026ndash;577. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/10548408.2024.2309180\u003c/span\u003e\u003cspan address=\"10.1080/10548408.2024.2309180\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDeng N, Li X (2018) Feeling a destination through the right photos: A machine learning model for DMOs\u0026rsquo; photo selection. Tour Manag 65:267\u0026ndash;278. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.tourman.2017.09.010\u003c/span\u003e\u003cspan address=\"10.1016/j.tourman.2017.09.010\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDixit SK, Prayag G (2022) Gastronomic tourism experiences and experiential marketing. Tourism Recreation Res 47(3):217\u0026ndash;220. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/02508281.2022.2065089\u003c/span\u003e\u003cspan address=\"10.1080/02508281.2022.2065089\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEngler J, Sikdar S, Lutz M, Strohmaier M (2023) SensePOLAR: Word sense aware interpretability for pre-trained contextual word embeddings. \u003cem\u003eArXiv Preprint ArXiv:2301.04704\u003c/em\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFellbaum C (2005) WordNet and wordnets. In: Brown, Keith (eds.), \u003cem\u003eEncyclopedia of Language and Linguistics\u003c/em\u003e, Second Edition, Oxford: Elsevier, 665\u0026ndash;670\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFilieri R, Yen DA, Yu Q (2021) #ILoveLondon: An exploration of the declaration of love towards a destination on Instagram. Tour Manag 85:104291. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.tourman.2021.104291\u003c/span\u003e\u003cspan address=\"10.1016/j.tourman.2021.104291\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFossgard K, Fredman P (2019) Dimensions in the nature-based tourism experiencescape: An explorative analysis. J Outdoor Recreation Tourism 28:100219. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jort.2019.04.001\u003c/span\u003e\u003cspan address=\"10.1016/j.jort.2019.04.001\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFu X, Liu X, Li Z (2024) Catching eyes of social media wanderers: How pictorial and textual cues in visitor-generated content shape users\u0026rsquo; cognitive-affective psychology. Tour Manag 100:104815\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGambetti A, Han Q (2022) Camera eats first: exploring food aesthetics portrayed on social media using deep learning. Int J Contemp Hospitality Manage 34(9):3300\u0026ndash;3331. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1108/IJCHM-09-2021-1206\u003c/span\u003e\u003cspan address=\"10.1108/IJCHM-09-2021-1206\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGon M (2020) Local experiences on Instagram: Social media data as source of evidence for experience design. J Destination Mark Manage 100435. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jdmm.2020.100435\u003c/span\u003e\u003cspan address=\"10.1016/j.jdmm.2020.100435\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGrootendorst M (2020) Bertopic: Leveraging bert and c-tf-idf to create easily interpretable topics. Version V0, 4\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHauser D, Leopold A, Egger R, Ganewita H, Herrgessell L (2022) Aesthetic perception analysis of destination pictures using #beautifuldestinations on Instagram. J Destination Mark Manage 24:100702. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jdmm.2022.100702\u003c/span\u003e\u003cspan address=\"10.1016/j.jdmm.2022.100702\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHe S, Liao W, Tavakoli HR, Yang M, Rosenhahn B, Pugeault N (2020) Image Captioning through Image Transformer. \u003cem\u003eProceedings of the Asian Conference on Computer Vision\u003c/em\u003e, 1\u0026ndash;17\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHong WCH, Ngan HFB, Yu J, Arbouw P (2023) Examining cultural differences in Airbnb naming convention and user reception: an eye-tracking study. J Travel Tourism Mark 40(6):475\u0026ndash;489. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/10548408.2023.2263764\u003c/span\u003e\u003cspan address=\"10.1080/10548408.2023.2263764\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHossain MZ, Sohel F, Shiratuddin MF, Laga H (2019) A comprehensive survey of deep learning for image captioning. ACM Comput Surv (CsUR) 51(6):1\u0026ndash;36\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHuang X, Kader MS, Kim S (2021) Let's play with emojis! How to make emojis more effective in social media advertising using promocodes and temporal orientation. Journal Res Interact Mark Adv online publication. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1108/JRIM-04-2021-0113\u003c/span\u003e\u003cspan address=\"10.1108/JRIM-04-2021-0113\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHuang Y, Chen CH, Khoo LP (2012) Products classification in emotional design using a basic-emotion based semantic differential method. Int J Ind Ergon 42(6):569\u0026ndash;580. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ergon.2012.09.002\u003c/span\u003e\u003cspan address=\"10.1016/j.ergon.2012.09.002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKiatkawsin K, Bui NA, Hrankai R, Jeong K (2021) The Moderating Roles of Sensation Seeking and Worry among Nature-Based Adventure Tourists. Int J Environ Res Public Health 18(4). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/ijerph18042021\u003c/span\u003e\u003cspan address=\"10.3390/ijerph18042021\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eK\u0026ouml;ltringer C, Dickinger A (2015) Analyzing destination branding and image from online sources: A web content mining approach. J Bus Res 68(9):1836\u0026ndash;1843\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKuo L, Chang T, Lai C\u0026ndash;C (2022) Multimedia webpage visual design and color emotion test. Multimedia Tools Appl 81(2):2621\u0026ndash;2636. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s11042-021-11684-4\u003c/span\u003e\u003cspan address=\"10.1007/s11042-021-11684-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLee Y\u0026ndash;K, Lee C\u0026ndash;K, Lee W, Ahmad MS (2021) Do hedonic and utilitarian values increase pro-environmental behavior and support for festivals? Asia Pac J Tourism Res 26(8):921\u0026ndash;934. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/10941665.2021.1927122\u003c/span\u003e\u003cspan address=\"10.1080/10941665.2021.1927122\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLeit\u0026atilde;o J, Pereira D, Gon\u0026ccedil;alves \u0026Acirc; (2021) Quality of Work Life and Contribution to Productivity: Assessing the Moderator Effects of Burnout Syndrome. Int J Environ Res Public Health 18(5). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/ijerph18052425\u003c/span\u003e\u003cspan address=\"10.3390/ijerph18052425\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLi H, Ji H, Liu H, Cai D, Gao H (2022) Is a picture worth a thousand words? Understanding the role of review photo sentiment and text-photo sentiment disparity using deep learning algorithms. Tour Manag 92:104559. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.tourman.2022.104559\u003c/span\u003e\u003cspan address=\"10.1016/j.tourman.2022.104559\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLi J, Abbasi A, Cheema A, Abraham LB (2020) Path to Purpose? How Online Customer Journeys Differ for Hedonic Versus Utilitarian Purchases. J Mark 84(4):127\u0026ndash;146. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/0022242920911628\u003c/span\u003e\u003cspan address=\"10.1177/0022242920911628\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLi Y, Xie Y (2020) Is a Picture Worth a Thousand Words? An Empirical Study of Image Content and Social Media Engagement. J Mark Res 57(1):1\u0026ndash;19. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/0022243719881113\u003c/span\u003e\u003cspan address=\"10.1177/0022243719881113\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLiadeli G, Sotgiu F, Verlegh PW (2023) A Meta-Analysis of the Effects of Brands\u0026rsquo; Owned Social Media on Social Media Engagement and Sales. J Mark 87(3):406\u0026ndash;427. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/00222429221123250\u003c/span\u003e\u003cspan address=\"10.1177/00222429221123250\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLim H, Childs M (2020) Visual storytelling on Instagram: branded photo narrative and the role of telepresence. J Res Interact Mark 14(1):33\u0026ndash;50. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1108/JRIM-09-2018-0115\u003c/span\u003e\u003cspan address=\"10.1108/JRIM-09-2018-0115\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLim JS, Choe M\u0026ndash;J, Zhang J, Noh G\u0026ndash;Y (2020) The role of wishful identification, emotional engagement, and parasocial relationships in repeated viewing of live-streaming games: A social cognitive theory perspective. Comput Hum Behav 108:106327. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.chb.2020.106327\u003c/span\u003e\u003cspan address=\"10.1016/j.chb.2020.106327\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLin F, Ryu K, Lee YK, Ali F (2024) Understanding the importance of caf\u0026eacute; portable cup design and environmental concerns on green on-the-go consumption: a comparative study of gender. J Travel Tourism Mark 41(9):1262\u0026ndash;1279. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/10548408.2024.2425101\u003c/span\u003e\u003cspan address=\"10.1080/10548408.2024.2425101\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLiu Q, Kusner MJ, Blunsom P (2020) A survey on contextual embeddings. \u003cem\u003earXiv preprint arXiv:2003.07278\u003c/em\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLiu M (2019) Correlation between Aesthetic Ability and Art Appreciation. Int J Trend Res Dev 6(5):14\u0026ndash;16\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eL\u0026oacute;pez-Chao V, Lopez-Pena V (2020) Aesthetical Appeal and Dissemination of Architectural Heritage Photographs in Instagram. Buildings 10(12):225. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/buildings10120225\u003c/span\u003e\u003cspan address=\"10.3390/buildings10120225\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMa Z, Palacios S (2021) Image-mining: exploring the impact of video content on the success of crowdfunding. J Mark Analytics 9(4):265\u0026ndash;285. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1057/s41270-021-00133-8\u003c/span\u003e\u003cspan address=\"10.1057/s41270-021-00133-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMathew B, Sikdar S, Lemmerich F, Strohmaier M (2020) The POLAR Framework: Polar Opposites Enable Interpretability of Pre-Trained Word Embeddings. In Y. Huang, I. King, T.-Y. Liu, \u0026amp; M. van Steen (Eds.), \u003cem\u003eProceedings of The Web Conference 2020\u003c/em\u003e (pp. 1548\u0026ndash;1558). ACM. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1145/3366423.3380227\u003c/span\u003e\u003cspan address=\"10.1145/3366423.3380227\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcInnes L, Healy J, Saul N, Gro\u0026szlig;berger L (2018) UMAP: Uniform Manifold Approximation and Projection. J Open Source Softw 3(29):861. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.21105/joss.00861\u003c/span\u003e\u003cspan address=\"10.21105/joss.00861\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMindak WA (1961) Fitting the semantic differential to the marketing problem. J Mark 25(4):28\u0026ndash;33\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNeuhofer B, Egger R, Yu J, Celuch K (2021) Designing experiences in the age of human transformation: An analysis of Burning Man. Annals Tourism Res 91:103310. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.annals.2021.103310\u003c/span\u003e\u003cspan address=\"10.1016/j.annals.2021.103310\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNguyen TK, Coustaty M, Guillaume JL (eds) (2018) \u003cem\u003eA new image segmentation approach based on the Louvain algorithm\u003c/em\u003e. IEEE. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://ieeexplore.ieee.org/servlet/opac?punumber=8491725\u003c/span\u003e\u003cspan address=\"http://ieeexplore.ieee.org/servlet/opac?punumber=8491725\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePicazo P, Moreno-Gil S (2019) Analysis of the projected image of tourism destinations on photographs: A literature review to prepare for the future. J Vacation Mark 25(1):3\u0026ndash;24. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/1356766717736350\u003c/span\u003e\u003cspan address=\"10.1177/1356766717736350\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRadford A, Kim JW, Hallacy C, Ramesh A, Goh G, Agarwal S, Sastry G, Askell A, Mishkin P, Clark J, Krueger G, Sutskever I (2021), February 26 \u003cem\u003eLearning Transferable Visual Models From Natural Language Supervision\u003c/em\u003e. http://arxiv.org/pdf/2103.00020v1\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eReelfs J, Mohaupt T, Sikdar S, Strohmaier M, Hohlfeld O (eds) (2022) Interpreting Emoji with Emoji. Association for Computational Linguistics\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eReimers N, Gurevych I (2019) Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks. \u003cem\u003eProceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP)\u003c/em\u003e, 3982\u0026ndash;3992, Association for Computational Linguistics\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRietveld R, van Dolen W, Mazloom M, Worring M (2020) What You Feel, Is What You Like Influence of Message Appeals on Customer Engagement on Instagram. J Interact Mark 49:20\u0026ndash;53. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.intmar.2019.06.003\u003c/span\u003e\u003cspan address=\"10.1016/j.intmar.2019.06.003\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRoggeveen AL, Grewal D, Townsend C, Krishnan R (2015) The Impact of Dynamic Presentation Format on Consumer Preferences for Hedonic Products and Services. J Mark 79(6):34\u0026ndash;49. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1509/jm.13.0521\u003c/span\u003e\u003cspan address=\"10.1509/jm.13.0521\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSoldat Z, Quintal V, Phau I (2024) Pursuing travel calculated hedonism: scale development and validation. J Travel Tourism Mark 41(5):760\u0026ndash;774. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/10548408.2024.2341653\u003c/span\u003e\u003cspan address=\"10.1080/10548408.2024.2341653\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eStead S, Wetzels R, Wetzels M, Odekerken-Schr\u0026ouml;der G, Mahr D (2022) Toward Multisensory Customer Experiences: A Cross-Disciplinary Bibliometric Review and Future Research Directions. J Service Res 109467052210799. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/10946705221079941\u003c/span\u003e\u003cspan address=\"10.1177/10946705221079941\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eStremersch S, Gonzalez J, Valenti A, Villanueva J (2023) The value of context-specific studies for marketing. J Acad Mark Sci 51(1):50\u0026ndash;65. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s11747-022-00872-9\u003c/span\u003e\u003cspan address=\"10.1007/s11747-022-00872-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSu L, Ye C, Huang Y (2024) Does destination nostalgic advertising enhance tourists\u0026rsquo; intentions to visit? The moderating role of destination type. Tour Manag 100:104810. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.tourman.2023.104810\u003c/span\u003e\u003cspan address=\"10.1016/j.tourman.2023.104810\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSung YH, Kim DH, Choi D, Lee SY (2020) Facebook ads not working in the same way: The effect of cultural orientation and message construals on consumer response to social media ads. Telematics Inform 52:101427. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.tele.2020.101427\u003c/span\u003e\u003cspan address=\"10.1016/j.tele.2020.101427\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTaberner I, Junc\u0026agrave; A (2021) Small-Scale Sport Events as Place Branding Platforms: A Content Analysis of Osona\u0026rsquo;s Projected Destination Image through Event-Related Pictures on Instagram. Sustainability 13(21):12255. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/su132112255\u003c/span\u003e\u003cspan address=\"10.3390/su132112255\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTaecharungroj V, Mathayomchan B (2021) Traveller-generated destination image: Analysing Flickr photos of 193 countries worldwide. Int J Tourism Res 23(3):417\u0026ndash;441\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTussyadiah IP (2014) Toward a Theoretical Foundation for Experience Design in Tourism. J Travel Res 53(5):543\u0026ndash;564. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/0047287513513172\u003c/span\u003e\u003cspan address=\"10.1177/0047287513513172\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVillamediana-Pedrosa JD, Vila-Lopez N, K\u0026uuml;ster-Boluda I (2018) Secrets to design an effective message on Facebook: an application to a touristic destination based on big data analysis. Curr Issues Tourism 13(6):1\u0026ndash;21. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/13683500.2018.1554625\u003c/span\u003e\u003cspan address=\"10.1080/13683500.2018.1554625\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVirtanen H, Bj\u0026ouml;rk P, Sj\u0026ouml;str\u0026ouml;m E (2017) Follow for follow: marketing of a start-up company on Instagram. J Small Bus Enterp Dev 24(3):468\u0026ndash;484\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVillamediana J, K\u0026uuml;ster I, Vila N (2019) Destination engagement on Facebook: Time and seasonality. Annals Tourism Res 79:102747\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVoss KE, Spangenberg ER, Grohmann B (2003) Measuring the hedonic and utilitarian dimensions of consumer attitude. J Mark Res 40(3):310\u0026ndash;320\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ede Vries L, Gensler S, Leeflang PS (2012) Popularity of Brand Posts on Brand Fan Pages: An Investigation of the Effects of Social Media Marketing. J Interact Mark 26(2):83\u0026ndash;91. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.intmar.2012.01.003\u003c/span\u003e\u003cspan address=\"10.1016/j.intmar.2012.01.003\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVu HQ, Song B, Li G, Law R (2025) Exploring emotional aspects of travel concepts via travel photos based on contrastive language-image pretraining. Tour Manag 108:105117\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang L, Luo Z (2025) Effects of tourists\u0026rsquo; online self-disclosure depth on readers\u0026rsquo; visit intentions: mediating roles of perceived social distance and perceived self-expressiveness. J Travel Tourism Mark 42(1):40\u0026ndash;64. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/10548408.2024.2425100\u003c/span\u003e\u003cspan address=\"10.1080/10548408.2024.2425100\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWies S, Bleier A, Edeling A (2023) Finding Goldilocks Influencers: How Follower Count Drives Social Media Engagement. J Mark 87(3):383\u0026ndash;405. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/00222429221125131\u003c/span\u003e\u003cspan address=\"10.1177/00222429221125131\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWu L, You JS (2022) Synchronous or Asynchronous Course: Business Students\u0026rsquo; Perspectives on an Optimized Modality of Online Teaching and Learning. J Mark Educ 44(2):265\u0026ndash;284. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/02734753221093740\u003c/span\u003e\u003cspan address=\"10.1177/02734753221093740\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eXiao X, Fang C, Lin H (2020) Characterizing Tourism Destination Image Using Photos\u0026rsquo; Visual Content. ISPRS Int J Geo-Information 9(12):730. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/ijgi9120730\u003c/span\u003e\u003cspan address=\"10.3390/ijgi9120730\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eXu J, Wang Y, Jiang Y (2023) How do social media tourist images influence destination attitudes? Effects of social comparison and envy. J Travel Tourism Mark 40(4):310\u0026ndash;325. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/10548408.2023.2245410\u003c/span\u003e\u003cspan address=\"10.1080/10548408.2023.2245410\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYou Q, Jin H, Wang Z, Fang C, Luo J (2016) Image captioning with semantic attention. In \u003cem\u003eProceedings of the IEEE conference on computer vision and pattern recognition\u003c/em\u003e (pp. 4651\u0026ndash;4659)\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYu J, Egger R (2021) Color and engagement in touristic Instagram pictures: A machine learning approach. Annals Tourism Res. 103204\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.annals.2021.103204\u003c/span\u003e\u003cspan address=\"10.1016/j.annals.2021.103204\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYu J, Hong WCH, Egger R (2025) The art of post captions: Readability and user engagement on social media. J Travel Res 64(4):853\u0026ndash;866\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYuda Bakti IGM, Rakhmawati T, Sumaedi S, Damayanti S (2020) Railway commuter line passengers\u0026rsquo; perceived service quality: hedonic and utilitarian framework. Transp Res Procedia 48:207\u0026ndash;217. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.trpro.2020.08.016\u003c/span\u003e\u003cspan address=\"10.1016/j.trpro.2020.08.016\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"information-technology-and-tourism","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jitt","sideBox":"Learn more about [Information Technology \u0026 Tourism](https://link.springer.com/journal/40558)","snPcode":"40558","submissionUrl":"https://submission.springernature.com/new-submission/40558/3","title":"Information Technology \u0026 Tourism","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"semantic differential, visual analysis, hedonic consumption, utilitarian consumption, machine learning","lastPublishedDoi":"10.21203/rs.3.rs-7591283/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7591283/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe pervasive rise of social media has placed visual content, especially photographs, at the forefront of digital marketing in tourism. As marketers strategically leverage emotionally resonant content on platforms, this research explores the relationship between hedonic and utilitarian dimensions embedded in pictorial representations and their impact on user engagement in tourism. By utilising a machine-generated semantic differential, this study analyses 33,001 picture-based posts from popular destinations on Instagram. The findings of this study reveal that user engagement dynamics vary across themes, with multimodal excursions (e.g., modes of transportation) emphasising informativeness over fun and thrill, while art and sculpture benefit from the integration of functionality. By using an interdisciplinary approach, this research explores the emotional connotations of visual content on the engagement of potential tourists. Overall, this paper underscores the importance of tailoring content strategies to specific thematic contexts and challenges preconceived notions about the universal appeal of certain pictorial elements.\u003c/p\u003e","manuscriptTitle":"Feeling travel photographs: Emotional labelling in machine-generated semantic differentials","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-15 08:30:41","doi":"10.21203/rs.3.rs-7591283/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-11-10T11:37:54+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-06T17:33:14+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-05T10:48:27+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"296914230762686658553576611076052018312","date":"2025-10-08T12:43:11+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"166012547691799494456782382213406578113","date":"2025-10-07T03:54:32+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-10-02T03:38:49+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-29T13:23:10+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-09-15T13:30:57+00:00","index":"","fulltext":""},{"type":"submitted","content":"Information Technology \u0026 Tourism","date":"2025-09-11T11:09:07+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"information-technology-and-tourism","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jitt","sideBox":"Learn more about [Information Technology \u0026 Tourism](https://link.springer.com/journal/40558)","snPcode":"40558","submissionUrl":"https://submission.springernature.com/new-submission/40558/3","title":"Information Technology \u0026 Tourism","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"11eedcfc-41cb-4c1c-8d0c-a681a8618763","owner":[],"postedDate":"October 15th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2026-05-04T16:01:41+00:00","versionOfRecord":{"articleIdentity":"rs-7591283","link":"https://doi.org/10.1007/s40558-026-00377-z","journal":{"identity":"information-technology-and-tourism","isVorOnly":false,"title":"Information Technology \u0026 Tourism"},"publishedOn":"2026-04-29 15:58:03","publishedOnDateReadable":"April 29th, 2026"},"versionCreatedAt":"2025-10-15 08:30:41","video":"","vorDoi":"10.1007/s40558-026-00377-z","vorDoiUrl":"https://doi.org/10.1007/s40558-026-00377-z","workflowStages":[]},"version":"v1","identity":"rs-7591283","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7591283","identity":"rs-7591283","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}