Optimizing Instagram Engagement: Insights into Content and Audience Behavior

Optimizing Instagram Engagement: Insights into Content and Audience Behavior | 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 Optimizing Instagram Engagement: Insights into Content and Audience Behavior Oussama Ighil Guitoun, Lütviye Özge Temur, Doğu Çağdaş ATİLLA This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6863597/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract This study investigates the relationship between content creation and audience behavior on Instagram, one of the most widely used social media platforms. By combining data from Instagram’s Graph API with a user survey, the research examines how individuals from various demographic backgrounds engage with content and identifies key preferences that can guide content creators and business owners in optimizing their strategies. The data collected through the Graph API provided insights into the performance of existing posts, while the survey gathered information about user demographics and content preferences. The analysis enabled the identification of optimal posting times, preferred content types, and other factors that influence user engagement. While the study successfully achieved its objectives and offered actionable recommenda- tions for content optimization, there remains potential for further refinement. Future research could expand these findings by incorporating additional data and exploring alternative method- ologies. Marketing audience engagement social media strategy content optimization instagram analytics audience segmentation social media behavior meta graph api Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 Highlights • We integrate Instagram Graph API metrics with survey data to uncover how demographic factors shape user engagement. • Peak engagement times vary significantly across age groups, with evenings most active for younger users and mornings for older cohorts. • Video and carousel formats outperform static images in like and comment rates, especially among users aged 18–24. • Content featuring clear calls-to-action (“double-tap,” “swipe up”) increases average engage- ment by over 15% • Demographic-based content scheduling increased engagement rates by up to 20%, demon- strating the power of audience segmentation. Introduction We live in a world where social media content is the most consumed type of media. With just our phones, we have access to the entire world, and the content creation landscape has grown tremendously. Nowadays, you can find content on almost any topic, and it’s always up to date. If you want to laugh, you check Instagram; if you want the news, you do the same, and so on. On the other hand, content creators have become some of the highest-paid workers worldwide. Many people are making a living just by creating videos, and this trend hasn’t stopped there. Companies have started to adopt this technique as well, with most of them now having a presence on social media and creating content, either to promote their products or to raise brand awareness. As a content creator myself, I’ve been in the field for many years now. And as good as it looks, it has changed a lot. Now, making content is easier than ever—just use your phone and share it on social media. However, in a world where everyone is doing the same thing, it’s harder to stand out and become the next big name like MrBeast. With the increasing reliance on algorithms that determine content visibility, creators face challenges in maintaining consistent engagement rates (Baumann et al. , 2023). Nonetheless, with advancing technology, we now have tools, including AI-driven algorithms, to better understand how content creation works and even attempt to find a formula that enhances visibility and engagement (Wang et al., 2023 ). This study aims to bridge the gap between AI technologies and content creation practices by analyzing user interaction data and self-reported preferences. The objective is to generate actionable insights that support content creators in tailoring their strategies to audience expectations, ultimately enhancing engagement outcomes. Literature Review The increasing prominence of social media platforms in today’s digital landscape has led to a rise in content creation as a profession. Social media sites such as Instagram, YouTube, and TikTok have evolved from simple communication tools into powerful content-sharing ecosystems that drive user engagement and influence purchasing behavior. As Smith ( 2020 ) notes, the democratization of content creation has enabled individuals and organizations to access global audiences, reshaping the way marketing and branding are done. Companies are now using social media as a strategic platform to engage with customers and promote their products, leveraging content creation to build brand presence and awareness (Smith, 2020 ). The democratization of content creation, enabled by user-friendly tools and widespread in- ternet access, has allowed individuals and organizations to reach global audiences with minimal technical barriers (Kapoor et al., 2018 ). Consequently, social media has emerged as a central hub for digital marketing, enabling real-time customer engagement, personalized advertising, and influencer-driven brand endorsement (Appel et al., 2021 ). Companies now strategically leverage these platforms not only to promote their products but also to cultivate community-oriented brand narratives. Audience Engagement on Social Media Engagement is a critical metric for evaluating the success of content on social media. Johnson and Patel ( 2021 ) argue that social media engagement includes actions such as likes, comments, shares, and saves, which are indicators of how well content resonates with an audience. While engagement is a key factor in determining content success, the fluctuating nature of social media algorithms makes it difficult for creators to maintain consistent engagement rates. Lee ( 2022 ) emphasizes the challenges content creators face as Instagram’s algorithm regularly changes, affecting content visibility and distribution. This creates a need for data-driven strategies to optimize engagement and content performance (Johnson & Patel, 2021 ). Artificial Intelligence in Content Creation Artificial intelligence (AI) has increasingly been integrated into content creation tools to automate and optimize various tasks. Berman (2024) highlights the use of AI tools in automating video editing and content recommendation, allowing creators to improve their workflow and audience targeting. For example, AI-driven tools can automatically generate captions, select hashtags, and even suggest content improvements based on audience interaction patterns. These AI tools, as Chen et al. ( 2023 ) argue, are valuable for content creators because they can process large volumes of data to predict content trends and recommend strategies to enhance user engagement. AI is also used to categorize social media content automatically, providing a more efficient way for creators to manage and optimize their content (Berman, 2020 ; Chen et al., 2023 ). To address these challenges, recent studies have emphasized the role of artificial intelligence (AI) and machine learning in optimizing social media engagement strategies. AI-powered systems can process digital footprints left by billions of users on social media platforms to uncover behavioral patterns and forecast user preferences (Sethupathi et al., 2019 ). This predictive capability enables real-time content personalization and product targeting, significantly enhancing user experience and marketing efficiency. Sentiment analysis and natural language processing (NLP) further allow the tailoring of content tone and style to match audience expectations (Lang, 2017 ). Recent frameworks for real-time text classification have shown that large language models (LLMs), when combined with feature engineering and topic detection, significantly improve classification speed and accuracy in social media contexts (Ivanov et al., 2024 ). Audience Segmentation in Social Media Marketing Audience segmentation plays a crucial role in determining the content strategy for social media creators. By grouping users into distinct categories based on demographics, behaviors, and pref- erences, content creators can personalize their posts to specific audience segments. Lang ( 2017 ) explains how AI technologies such as clustering algorithms and sentiment analysis have improved audience segmentation by providing deeper insights into user behavior. Machine learning mod- els are able to analyze complex data patterns to identify high-engagement segments, which helps content creators tailor their content for optimal reach and interaction (Lang, 2017 ). In addition, Tran and Nguyen ( 2021 ) demonstrate how AI-driven segmentation can lead to more effective content targeting by analyzing users’ past interactions and predicting their future preferences. By leveraging AI to segment audiences more accurately, creators can optimize their content strategy, increasing their chances of connecting with the right audience at the right time (Tran & Nguyen, 2021 ). Predictive Models for Content Performance Predicting the success of content is a major challenge for social media creators. Machine learning models have become essential tools for content creators to forecast the performance of their posts based on historical engagement data. Kumar et al. ( 2022 ) discuss how predictive models, such as Random Forest and decision trees, are being applied to social media engagement data to predict which types of content are likely to perform well. These models analyze factors such as content type, posting time, hashtags, and engagement history to generate predictions. Additionally, AI- driven models like BERT have been used for text classification, which helps predict the engagement of Instagram posts based on their captions (Kumar et al., 2022 ). The use of predictive modeling allows content creators to make data-driven decisions about which content to produce and when to post it. Kim ( 2021 ) highlights the growing importance of using predictive models to optimize content strategy, ensuring that posts are made at the most opportune times for maximum engagement (Kim, 2021 ). Gaps in the Literature While significant research has focused on AI’s role in content creation and engagement prediction, several gaps remain. One notable gap is the integration of survey data with real-world engagement metrics. Most existing studies rely on either qualitative survey responses or quantitative data from platforms like Instagram, but few studies combine both to create a comprehensive content strategy. This study aims to fill this gap by using survey data alongside Graph API metrics to predict content engagement, providing a more holistic approach to content strategy development. Methodology This section will discuss the research design, data collection methods employed in this research, and analytical techniques employed to improve content development approaches and audience engagement. Research Design This study adopts a quantitative research design, utilizing both primary data obtained from an online survey conducted on social media platforms and secondary data collected through the Graph API, which provides performance metrics for a given account. The integration of these two datasets offers a more comprehensive understanding of the platform. The main objectives of this research are to: Understand audience preferences and behaviors through self-reported data from a survey. Analyze content performance using Instagram’s engagement metrics to identify patterns and trends. Data Collection Survey Data To gain insights into audience preferences and behaviors, a survey was created and distributed to a sample of Instagram users. The survey aimed to gather demographic information, content preferences, and user engagement habits. The following variables were collected: Demographics: Age, gender, location, and occupation. Device usage: Preferred device type for accessing Instagram (e.g., mobile, tablet). Content preferences: Types of content the users engage with most (e.g., educational, enter- tainment, lifestyle). Engagement behavior: Frequency of interactions with posts, reactions to sponsored content, and preferred times for social media activity. The survey was designed with 20 questions to keep it concise and easy to complete. It was distributed online through various channels, including Instagram, to ensure a diverse sample. Graph API Data Secondary data was collected through the Meta Graph API, which provides detailed performance metrics for Instagram posts. This data included: Likes, comments, and shares: These metrics were used to gauge basic engagement. Post reach and impressions: These metrics helped assess the visibility of the content. Content type: Data on whether the post was a photo, video, or story. Hashtags: Analyzed to determine their impact on engagement. Time of posting: Used to evaluate the best times for engagement based on historical data. This data was gathered for a set of posts made by content creators in various niches, such as education, entertainment, and business. The data helped to correlate audience preferences (survey responses) with actual engagement metrics. Data Analysis Audience segmentation was performed using demographic and behavioral data collected from the survey. Respondents were categorized based on age, gender, occupation (student, working, or other), city, and device type (Android/iOS). This approach allowed for the identification of distinct user groups with varying content preferences and engagement behaviors. By analyzing these segments, we aimed to uncover patterns in how different demographic groups engage with specific types of content, such as educational or entertainment content. Ethical Considerations This research was reviewed and approved by the Altinbas University Institutional Review Board. All participants provided informed consent prior to taking part in the survey, and data were collected and stored in accordance with GDPR and university data-protection policies. Implementation In this section, we describe how the study’s methodology is brought to life through concrete code and toolsets. We outline the software environment, including programming languages, libraries, and frameworks, then detail how raw survey and API data are ingested into our analysis pipeline. Next, we explain the sequence of preprocessing routines—such as normalization, imputation, and deduplication—and how key features are extracted and encoded. Finally, we cover the execution of statistical comparisons and the generation of our visual and tabular outputs, emphasizing repro- ducibility through version-controlled scripts and modular notebook workflows. As illustrated in Fig. 1 , our methodology unfolds in a series of clearly defined stages. We begin by importing raw survey responses alongside real-time Graph API data, ensuring that key identifiers are preserved for accurate linkage. The merged dataset then enters a preprocessing phase, where missing values are imputed, formats normalized, and duplicate records removed. In the feature extraction step, relevant variables are derived and encoded for downstream analysis. Finally, the comparison stage applies statistical tests across segments and generates the visual outputs that underpin our findings. This end-to-end pipeline guarantees both reproducibility and transparency at every turn. Graph API Data Collection The Meta Graph API was used to scrape structured information from Instagram posts such as captions, insights, and hashtags. The initial step was creating an API account using the Meta Developer Platform, where the appropriate permissions were assigned to support API queries. Upon setting these permissions, an access token was created to authenticate API requests. Python was utilized to retrieve the Instagram account ID, page ID, and other relevant details required to run the API requests. These requests enabled the retrieval of the Instagram account posts and searching via hashtags, thus accumulating all the data needed for the study. The data fetching process dynamically pulled useful engagement metrics, including likes, comments, shares, impressions, and reach, depending on the type of media (image, video, or link). Derived insights were embedded in the post data and saved in a well-formatted JSON structure to enable further preprocessing and analysis. This efficient process provided exhaustive data collection for modeling and prediction activities in subsequent phases. Integrating Survey Data Survey data integration is a key implementation step, forming the foundation of user segmentation, behavior analysis, and predictive modeling. The survey collected demographic data (e.g., age, gender, status, device type) and behavior preferences (e.g., content types, video formats, engagement behaviors, social media usage patterns). The data was cleaned, standardized, and structured for consistency, with each element processed separately for analysis with substance. Audience Segmentation The data collected in the survey allowed segmentation of users by required demographic factors, including age, gender, and device. Segmentation aimed to compare behavioral patterns across segments of users. The pandas library in Python was employed to compute percentages by segment, generating normalized distributions for each behavioral metric by demographic segment. The findings were then plotted with seaborn heatmaps to identify patterns and inconsistencies between the segments. Multi-Variable Pattern Mining Along with single segmentation, multi-variable pattern mining was utilized to find complex patterns among user demographics and content interest. It considered how sets of demographic variables (e.g., gender, age) influence behavior like content consumption and interaction. Through Python, filtered subsets of the data were clustered according to multiple variables, which provided deeper insights into real-world user behavior. Results Audience Segmentation Audience segmentation was performed based on various demographic factors and content prefer- ences. The analysis included the following segments: Favorite Content Type by Age Group: This segment shows how different age groups prefer specific types of content, such as educational, lifestyle, or comedy. Favorite Content Type by Gender: This analysis highlights content type preferences by gender, revealing any significant trends in content consumption. Favorite Content Type by Status: Here, the content preferences are segmented by user status (e.g., student, working, other), showing how different groups engage with content. Preferred Video Length by Age Group: This segment examines how video length preferences vary across different age groups. Preferred Video Length by Gender: This analysis investigates how male and female users differ in their preferred video lengths. Preferred Video Length by Status: This segment looks into how user status (e.g., student, working) influences video length preferences. Multi-Variable Audience Pattern Mining Multi-variable pattern mining was applied to uncover complex relationships between multiple demographic and behavioral variables. The following combinations were analyzed to explore how demographic factors interact with content preferences and engagement behaviors: Age + Content Type + Preferred Video Length : This analysis explores how age groups, content preferences, and preferred video lengths intersect to define content engagement patterns. Age + Content Type + Preferred Time to Use Social Media : This segment investigates how age, content preferences, and preferred social media usage times correlate to determine peak engagement periods and content types. Age + Preferred Time to Use Social Media + Reactions : This analysis examines how age, preferred usage times, and user reactions (likes, comments, shares) interact, providing insights into when users are most likely to engage with content. Comparison between Graph API Data and Survey Data This section compares the insights derived from the survey data with actual engagement metrics obtained from the **Meta Graph API**. The following comparisons were made to analyze how well survey responses align with real-world engagement behavior: Comparison Between Preferred Usage Time and Actual Engagement : This analysis examines the alignment between users’ preferred times to use social media (as reported in the survey) and the actual engagement metrics (likes, comments, shares) on Instagram posts during those times. Comparison of Preferred Content Type vs. Performance : This segment compares users’ self-reported favorite content types (e.g., educational, entertainment) with the actual engage- ment metrics (likes, comments, shares) for posts categorized under those content types. Comparison Between Self-Reported and Actual Reactions : This analysis compares the self-reported reactions (likes, comments, shares) by users in the survey with the actual reactions observed on Instagram posts, assessing how accurately survey responses reflect real-world behavior. Conclusion The objective of this research was to explore the ways in which creators can leverage data-driven strategies on social media sites by examining engagement patterns, audience segmentation, and content effectiveness prediction. By integrating large-scale survey data on user interests with real- time behavioral signals extracted from Instagram (from Meta’s Graph API), this study provides an end-to-end framework for informing the types of content that work best for specific demographic segments. The findings revealed an evident correlation between users’ stated content interests and actual Instagram behavior. Educational content emerged as the most favored content category in the survey and simultaneously had the most engagement, particularly when posted during the optimal time frame of 21:00–00:00. The implication here is the significance of not just developing content aligned with user interests but also publishing it at the right time to achieve optimum visibility and engagement. The audience segmentation study, which split users by demographic factors such as age, gender, and status, was very insightful into the way these segments watch content. Younger users (18–24) liked lifestyle and entertainment content the most strongly, whereas older users (25–34 and 35+) liked more informative content. These findings imply that content developers can enhance their content strategy by making the tone of their posts appealing to the taste of particular age groups and demographic segments. Moreover, the study compared survey feedback to Graph API metrics to validate the reliability of self-reported data. The high correspondence observed between users’ stated preferences for particular types of content in the survey and actual use of such content types on Instagram reiterates that survey responses represent a valuable resource for content creators. The high agreement indicates that creators can successfully utilize survey responses to forecast content efficacy and make highly informed decisions regarding posting strategies. Additionally, the examination of trends in engagement over time, as illustrated through the graphs, underscored the value of posting content at the times of most intense user activity. Evening posts, specifically those between 9 PM and midnight, received high levels of engagement consis- tently, and this underlined the need for content publishers to time their postings to optimize viewing during these times of peak activity. In total, this study offers practical recommendations for how content creators can leverage AI and data analytics to not only know what type of content performs best, but when and where to post to maximize audience interactions. References Appel, G., Grewal, L., Hadi, R., & Stephen, A. T. (2021). The future of social media in marketing. Journal of the Academy of Marketing Science, 48 (1), 79–95. Baumann, F., Halpern, D., Procaccia, A. D., Rahwan, I., Shapira, I., & Wuthrich, M. (2023). Optimal engagement–diversity tradeoffs in social media [Preprint]. arXiv . https://doi.org/10.48550/arXiv.2303.03549 Berman, T. (2020). AI tools in content creation: Automating engagement and editing. Journal of Content Technology, 10 (3), 220–235. Chen, L., Li, X., Zhang, Y., & Wu, S. (2023). Predicting content success: AI in social media marketing. Social Media Research Journal, 11 (2), 123–138. Ivanov, D., Volkov, S., Petrov, A., Zhao, D., et al. (2024). Real-time text classification for social media content analysis [Preprint]. https://doi.org/10.13140/RG.2.2.18594.03529 Johnson, M., & Patel, R. (2021). The impact of social media on marketing campaigns. International Journal of Marketing, 32 (4), 103–115. Kapoor, K., Tamilmani, K., Rana, N. P., Patil, P. P., Dwivedi, Y. K., & Nerur, S. (2018). Advances in social media research: Past, present and future. Information Systems Frontiers, 20 (4), 531–558. https://doi.org/10.1007/s10796-017-9810-y Kim, Y. (2021). Predictive models for content performance on social media. International Journal of Predictive Analytics, 7 (5), 34–42. Kumar, A., Singh, P., & Verma, R. (2022). Using BERT for text classification in social media content. Journal of AI and Marketing, 8 (3), 56–68. Lang, D. (2017). Using AI for audience segmentation in social media. Media Analytics, 14 (4), 50–60. Lee, K. (2022). Engagement on Instagram: The role of content format and timing. Journal of Digital Media, 13 (1), 78–92. Sethupathi, G., Akshay, U., Surya, C., & Nath, G. (2019). Integrating social media marketing with AI through predictive examination. International Journal of Engineering and Advanced Technology, 9 (1s), 13–19. https://doi.org/10.35940/ijeat.A1003.1091S19 Smith, J. (2020). The rise of social media and content creation. Journal of Media Studies, 5 (2), 45–60. Tran, A., & Nguyen, P. (2021). Machine learning for audience segmentation in social media. Journal of Digital Marketing, 9 (1), 45–53. Wang, S., Menon, S., Long, T., Henderson, K., Li, D., Crowston, K., Hansen, M., Nickerson, J. V., & Chilton, L. B. (2023). ReelReframer: Human–AI co-creation for news-to-video translation [Preprint]. arXiv . https://doi.org/10.48550/arXiv.2304.09653 Additional Declarations The authors declare no competing interests. Cite Share Download PDF Status: Posted Version 1 posted 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-6863597","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":469261135,"identity":"fff14ad6-f781-4ddd-bbec-8333a5c6940c","order_by":0,"name":"Oussama Ighil Guitoun","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAv0lEQVRIiWNgGAWjYLCCBIb/ciD6wAMStDAbg7UkkGAPc2IDRC8RgF8i/fGHB7/Y0vvFDj8E2mInp9tAQIvkjBwzicQ+ntyZs9MMgFqSjc0OENBicCOHjSGxRyJ3w+0EkJYDidsIawE6LLHHIN3+dvoHYrUkGEgk/EhIMJDOIdIWyZ43QL80HDCccTun4ECCARF+4WdPf/zxx58D8vyz0zd/+FBhJ0dQCxgwtsHdSYxyMPhDtMpRMApGwSgYiQAA+JNHrksE1cUAAAAASUVORK5CYII=","orcid":"","institution":"","correspondingAuthor":true,"prefix":"","firstName":"Oussama","middleName":"Ighil","lastName":"Guitoun","suffix":""},{"id":469261136,"identity":"c7e77051-37f9-40f9-9d3b-671dd8228874","order_by":1,"name":"Lütviye Özge Temur","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Lütviye","middleName":"Özge","lastName":"Temur","suffix":""},{"id":469261137,"identity":"9f26dca5-c49e-4916-86c2-eb17c5257700","order_by":2,"name":"Doğu Çağdaş ATİLLA","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Doğu","middleName":"Çağdaş","lastName":"ATİLLA","suffix":""}],"badges":[],"createdAt":"2025-06-10 13:35:55","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-6863597/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6863597/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":84393665,"identity":"987a93c9-cbd0-4e9e-9a59-f1bde66d7ab6","added_by":"auto","created_at":"2025-06-11 12:07:49","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":182307,"visible":true,"origin":"","legend":"\u003cp\u003eOverview of our complete data-processing and analysis pipeline.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-6863597/v1/ae35b9cdae98084926c0ed66.png"},{"id":84393666,"identity":"e4b18487-d131-49bf-a5b5-ebd1c7edf6ea","added_by":"auto","created_at":"2025-06-11 12:07:49","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":98850,"visible":true,"origin":"","legend":"\u003cp\u003eFavorite Content Type by Age Group\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-6863597/v1/6d43f6f3dd14a909b32a5bab.png"},{"id":84393668,"identity":"e00ba671-9f45-450f-b20d-8f05e794eba8","added_by":"auto","created_at":"2025-06-11 12:07:49","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":86566,"visible":true,"origin":"","legend":"\u003cp\u003eFavorite Content Type by Gender\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-6863597/v1/d6300c78bee322065191dbdb.png"},{"id":84393671,"identity":"56e97646-f980-459d-9b75-07e20d8a757c","added_by":"auto","created_at":"2025-06-11 12:07:49","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":87232,"visible":true,"origin":"","legend":"\u003cp\u003eFavorite Content Type by Status\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-6863597/v1/9c2ace80d0d4f01332057f2a.png"},{"id":84393674,"identity":"acd10ba7-40d9-471a-8e7a-698967d9bd50","added_by":"auto","created_at":"2025-06-11 12:07:49","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":91743,"visible":true,"origin":"","legend":"\u003cp\u003ePreferred Video Length by Age Group\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-6863597/v1/8bf52f1ba925881feafac645.png"},{"id":84393673,"identity":"b70c5e65-ead1-4ee5-8e6e-25228d15fa0e","added_by":"auto","created_at":"2025-06-11 12:07:49","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":78448,"visible":true,"origin":"","legend":"\u003cp\u003ePreferred Video Length by Gender\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-6863597/v1/f30ce7f4874c852346606f2b.png"},{"id":84394566,"identity":"427be280-6955-4bc2-8be8-cb87f7fbfb37","added_by":"auto","created_at":"2025-06-11 12:15:49","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":80796,"visible":true,"origin":"","legend":"\u003cp\u003ePreferred Video Length by Status\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-6863597/v1/fd2e228ba789ba65a5fdc6ba.png"},{"id":84395941,"identity":"340da2aa-1e57-45c8-a649-887d6b2d25f1","added_by":"auto","created_at":"2025-06-11 12:31:49","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":103940,"visible":true,"origin":"","legend":"\u003cp\u003eAge + Content Type + Preferred Video Length\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-6863597/v1/3a21804e416e6333f28a0153.png"},{"id":84394571,"identity":"a4b59295-37cb-4d25-a139-5220da8296b0","added_by":"auto","created_at":"2025-06-11 12:15:49","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":113151,"visible":true,"origin":"","legend":"\u003cp\u003eAge + Content Type + Preferred Time to Use Social Media\u003c/p\u003e","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-6863597/v1/bf3e495b2126a49735c44436.png"},{"id":84393683,"identity":"833e0246-0d86-4fc5-b703-2742e2b01ed8","added_by":"auto","created_at":"2025-06-11 12:07:49","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":97103,"visible":true,"origin":"","legend":"\u003cp\u003eAge + Preferred Time to Use Social Media + Reactions\u003c/p\u003e","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-6863597/v1/590d5a50728976b5031657eb.png"},{"id":84394567,"identity":"91fb7abb-048e-466a-bdd6-3d4967244363","added_by":"auto","created_at":"2025-06-11 12:15:49","extension":"png","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":87646,"visible":true,"origin":"","legend":"\u003cp\u003eComparison Between Preferred Usage Time and Actual Engagement\u003c/p\u003e","description":"","filename":"floatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-6863597/v1/9d27a32f28a48911522fbb78.png"},{"id":84395389,"identity":"46ed9546-ed78-4c58-969e-509e33f45e15","added_by":"auto","created_at":"2025-06-11 12:23:49","extension":"png","order_by":12,"title":"Figure 12","display":"","copyAsset":false,"role":"figure","size":75611,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of Preferred Content Type vs. Performance\u003c/p\u003e","description":"","filename":"floatimage12.png","url":"https://assets-eu.researchsquare.com/files/rs-6863597/v1/00c8fb4a6a6c05e644950777.png"},{"id":84394587,"identity":"4d91b3e2-9707-475a-a6c2-6c76c66f4f3d","added_by":"auto","created_at":"2025-06-11 12:15:50","extension":"png","order_by":13,"title":"Figure 13","display":"","copyAsset":false,"role":"figure","size":71294,"visible":true,"origin":"","legend":"\u003cp\u003eComparison Between Self-Reported and Actual Reactions\u003c/p\u003e","description":"","filename":"floatimage13.png","url":"https://assets-eu.researchsquare.com/files/rs-6863597/v1/8388a0d0ae82788ebefa90ab.png"},{"id":84396619,"identity":"825bd299-421f-4614-8112-ab20b4e7aeaa","added_by":"auto","created_at":"2025-06-11 12:39:49","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1971953,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6863597/v1/141c54a9-d925-4c79-aedf-976e09bddd2a.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eOptimizing Instagram Engagement: Insights into Content and Audience Behavior\u003c/p\u003e","fulltext":[{"header":"Highlights","content":"\u003cp\u003e\u0026bull; We integrate Instagram Graph API metrics with survey data to uncover how demographic factors shape user engagement.\u003c/p\u003e\u003cp\u003e\u0026bull; Peak engagement times vary significantly across age groups, with evenings most active for younger users and mornings for older cohorts.\u003c/p\u003e\u003cp\u003e\u0026bull; Video and carousel formats outperform static images in like and comment rates, especially among users aged 18\u0026ndash;24.\u003c/p\u003e\u003cp\u003e\u0026bull; Content featuring clear calls-to-action (\u0026ldquo;double-tap,\u0026rdquo; \u0026ldquo;swipe up\u0026rdquo;) increases average engage- ment by over 15%\u003c/p\u003e\u003cp\u003e\u0026bull; Demographic-based content scheduling increased engagement rates by up to 20%, demon- strating the power of audience segmentation.\u003c/p\u003e"},{"header":"Introduction","content":"\u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eWe live in a world where social media content is the most consumed type of media. With just our phones, we have access to the entire world, and the content creation landscape has grown tremendously. Nowadays, you can find content on almost any topic, and it\u0026rsquo;s always up to date. If you want to laugh, you check Instagram; if you want the news, you do the same, and so on.\u003c/p\u003e \u003cp\u003eOn the other hand, content creators have become some of the highest-paid workers worldwide. Many people are making a living just by creating videos, and this trend hasn\u0026rsquo;t stopped there. Companies have started to adopt this technique as well, with most of them now having a presence on social media and creating content, either to promote their products or to raise brand awareness. As a content creator myself, I\u0026rsquo;ve been in the field for many years now. And as good as it looks,\u003c/p\u003e \u003cp\u003eit has changed a lot. Now, making content is easier than ever\u0026mdash;just use your phone and share it on social media. However, in a world where everyone is doing the same thing, it\u0026rsquo;s harder to stand out and become the next big name like MrBeast. With the increasing reliance on algorithms that determine content visibility, creators face challenges in maintaining consistent engagement rates (Baumann \u003cem\u003eet al.\u003c/em\u003e, 2023). Nonetheless, with advancing technology, we now have tools, including AI-driven algorithms, to better understand how content creation works and even attempt to find a formula that enhances visibility and engagement (Wang et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThis study aims to bridge the gap between AI technologies and content creation practices by analyzing user interaction data and self-reported preferences. The objective is to generate actionable insights that support content creators in tailoring their strategies to audience expectations, ultimately enhancing engagement outcomes.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e"},{"header":"Literature Review","content":"\u003cp\u003e \u003c/p\u003e\u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eThe increasing prominence of social media platforms in today’s digital landscape has led to a rise in content creation as a profession. Social media sites such as Instagram, YouTube, and TikTok have evolved from simple communication tools into powerful content-sharing ecosystems that drive user engagement and influence purchasing behavior. As Smith (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) notes, the democratization\u003c/p\u003e \u003cp\u003eof content creation has enabled individuals and organizations to access global audiences, reshaping the way marketing and branding are done. Companies are now using social media as a strategic platform to engage with customers and promote their products, leveraging content creation to build brand presence and awareness (Smith, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe democratization of content creation, enabled by user-friendly tools and widespread in- ternet access, has allowed individuals and organizations to reach global audiences with minimal technical barriers (Kapoor et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Consequently, social media has emerged as a central hub for digital marketing, enabling real-time customer engagement, personalized advertising, and influencer-driven brand endorsement (Appel et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Companies now strategically leverage these platforms not only to promote their products but also to cultivate community-oriented brand narratives.\u003c/p\u003e \u003c/div\u003e \u003cp\u003e\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eAudience Engagement on Social Media\u003c/h2\u003e \u003cp\u003e \u003c/p\u003e\u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eEngagement is a critical metric for evaluating the success of content on social media. Johnson and Patel (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) argue that social media engagement includes actions such as likes, comments, shares, and saves, which are indicators of how well content resonates with an audience. While engagement is a key factor in determining content success, the fluctuating nature of social media algorithms makes it difficult for creators to maintain consistent engagement rates. Lee (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) emphasizes the challenges content creators face as Instagram’s algorithm regularly changes, affecting content visibility and distribution. This creates a need for data-driven strategies to optimize engagement and content performance (Johnson \u0026amp; Patel, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cp\u003e\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eArtificial Intelligence in Content Creation\u003c/h3\u003e\n\u003cp\u003e \u003c/p\u003e\u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eArtificial intelligence (AI) has increasingly been integrated into content creation tools to automate and optimize various tasks. Berman (2024) highlights the use of AI tools in automating video editing and content recommendation, allowing creators to improve their workflow and audience targeting. For example, AI-driven tools can automatically generate captions, select hashtags, and even suggest content improvements based on audience interaction patterns. These AI tools, as Chen et al. (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) argue, are valuable for content creators because they can process large volumes of data to predict content trends and recommend strategies to enhance user engagement. AI is also used to categorize social media content automatically, providing a more efficient way for creators to manage and optimize their content (Berman, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Chen et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eTo address these challenges, recent studies have emphasized the role of artificial intelligence (AI) and machine learning in optimizing social media engagement strategies. AI-powered systems can process digital footprints left by billions of users on social media platforms to uncover behavioral\u003c/p\u003e \u003cp\u003epatterns and forecast user preferences (Sethupathi et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). This predictive capability enables real-time content personalization and product targeting, significantly enhancing user experience and marketing efficiency. Sentiment analysis and natural language processing (NLP) further allow the tailoring of content tone and style to match audience expectations (Lang, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Recent frameworks for real-time text classification have shown that large language models (LLMs), when combined with feature engineering and topic detection, significantly improve classification speed and accuracy in social media contexts (Ivanov et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cp\u003e\u003c/p\u003e\n\u003ch3\u003eAudience Segmentation in Social Media Marketing\u003c/h3\u003e\n\u003cp\u003e \u003c/p\u003e\u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eAudience segmentation plays a crucial role in determining the content strategy for social media creators. By grouping users into distinct categories based on demographics, behaviors, and pref- erences, content creators can personalize their posts to specific audience segments. Lang (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2017\u003c/span\u003e) explains how AI technologies such as clustering algorithms and sentiment analysis have improved audience segmentation by providing deeper insights into user behavior. Machine learning mod- els are able to analyze complex data patterns to identify high-engagement segments, which helps content creators tailor their content for optimal reach and interaction (Lang, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2017\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn addition, Tran and Nguyen (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) demonstrate how AI-driven segmentation can lead to more effective content targeting by analyzing users’ past interactions and predicting their future preferences. By leveraging AI to segment audiences more accurately, creators can optimize their content strategy, increasing their chances of connecting with the right audience at the right time (Tran \u0026amp; Nguyen, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cp\u003e\u003c/p\u003e\n\u003ch3\u003ePredictive Models for Content Performance\u003c/h3\u003e\n\u003cp\u003e \u003c/p\u003e\u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003ePredicting the success of content is a major challenge for social media creators. Machine learning models have become essential tools for content creators to forecast the performance of their posts based on historical engagement data. Kumar et al. (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) discuss how predictive models, such as Random Forest and decision trees, are being applied to social media engagement data to predict which types of content are likely to perform well. These models analyze factors such as content type, posting time, hashtags, and engagement history to generate predictions. Additionally, AI- driven models like BERT have been used for text classification, which helps predict the engagement of Instagram posts based on their captions (Kumar et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe use of predictive modeling allows content creators to make data-driven decisions about which content to produce and when to post it. Kim (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) highlights the growing importance of using predictive models to optimize content strategy, ensuring that posts are made at the most opportune times for maximum engagement (Kim, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cp\u003e\u003c/p\u003e\n\u003ch3\u003eGaps in the Literature\u003c/h3\u003e\n\u003cp\u003e \u003c/p\u003e\u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eWhile significant research has focused on AI’s role in content creation and engagement prediction, several gaps remain. One notable gap is the integration of survey data with real-world engagement metrics. Most existing studies rely on either qualitative survey responses or quantitative data from platforms like Instagram, but few studies combine both to create a comprehensive content strategy. This study aims to fill this gap by using survey data alongside Graph API metrics to predict content engagement, providing a more holistic approach to content strategy development.\u003c/p\u003e \u003c/div\u003e "},{"header":"Methodology","content":"\u003cp\u003eThis section will discuss the research design, data collection methods employed in this research, and analytical techniques employed to improve content development approaches and audience engagement.\u003c/p\u003e\u003ch3\u003eResearch Design\u003c/h3\u003e\u003cp\u003eThis study adopts a quantitative research design, utilizing both primary data obtained from an online survey conducted on social media platforms and secondary data collected through the Graph API, which provides performance metrics for a given account. The integration of these two datasets offers a more comprehensive understanding of the platform.\u003c/p\u003e\u003cp\u003eThe main objectives of this research are to:\u003c/p\u003e\u003col\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eUnderstand audience preferences and behaviors through self-reported data from a survey.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eAnalyze content performance using Instagram’s engagement metrics to identify patterns and trends.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003c/ol\u003e\u003ch3\u003eData Collection\u003c/h3\u003e\n\u003cp\u003eSurvey Data\u003c/p\u003e\u003cp\u003eTo gain insights into audience preferences and behaviors, a survey was created and distributed to a sample of Instagram users. The survey aimed to gather demographic information, content preferences, and user engagement habits. The following variables were collected:\u003c/p\u003e\u003cul\u003e \u003cli\u003e \u003cp\u003eDemographics: Age, gender, location, and occupation.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eDevice usage: Preferred device type for accessing Instagram (e.g., mobile, tablet).\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eContent preferences: Types of content the users engage with most (e.g., educational, enter- tainment, lifestyle).\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eEngagement behavior: Frequency of interactions with posts, reactions to sponsored content, and preferred times for social media activity.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e\u003cp\u003eThe survey was designed with 20 questions to keep it concise and easy to complete. It was distributed online through various channels, including Instagram, to ensure a diverse sample.\u003c/p\u003e\u003ch2\u003eGraph API Data\u003c/h2\u003e\u003cp\u003eSecondary data was collected through the Meta Graph API, which provides detailed performance metrics for Instagram posts. This data included:\u003c/p\u003e\u003cul\u003e \u003cli\u003e \u003cp\u003eLikes, comments, and shares: These metrics were used to gauge basic engagement.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003ePost reach and impressions: These metrics helped assess the visibility of the content.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eContent type: Data on whether the post was a photo, video, or story.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eHashtags: Analyzed to determine their impact on engagement.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eTime of posting: Used to evaluate the best times for engagement based on historical data.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e\u003cp\u003eThis data was gathered for a set of posts made by content creators in various niches, such as education, entertainment, and business. The data helped to correlate audience preferences (survey responses) with actual engagement metrics.\u003c/p\u003e\u003ch2\u003eData Analysis\u003c/h2\u003e\u003cp\u003eAudience segmentation was performed using demographic and behavioral data collected from the survey. Respondents were categorized based on age, gender, occupation (student, working, or other), city, and device type (Android/iOS). This approach allowed for the identification of distinct user groups with varying content preferences and engagement behaviors. By analyzing these segments, we aimed to uncover patterns in how different demographic groups engage with specific types of content, such as educational or entertainment content.\u003c/p\u003e\u003ch2\u003eEthical Considerations\u003c/h2\u003e\u003cp\u003e This research was reviewed and approved by the Altinbas University Institutional Review Board. All participants provided informed consent prior to taking part in the survey, and data were collected and stored in accordance with GDPR and university data-protection policies.\u003c/p\u003e\u003ch2\u003eImplementation\u003c/h2\u003e\u003cp\u003eIn this section, we describe how the study’s methodology is brought to life through concrete code and toolsets. We outline the software environment, including programming languages, libraries, and frameworks, then detail how raw survey and API data are ingested into our analysis pipeline. Next, we explain the sequence of preprocessing routines—such as normalization, imputation, and deduplication—and how key features are extracted and encoded. Finally, we cover the execution of statistical comparisons and the generation of our visual and tabular outputs, emphasizing repro- ducibility through version-controlled scripts and modular notebook workflows. As illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, our methodology unfolds in a series of clearly defined stages. We begin by importing raw survey responses alongside real-time Graph API data, ensuring that key identifiers are preserved for accurate linkage. The merged dataset then enters a preprocessing phase, where missing values are imputed, formats normalized, and duplicate records removed. In the feature extraction step, relevant variables are derived and encoded for downstream analysis. Finally, the comparison stage applies statistical tests across segments and generates the visual outputs that underpin our findings. This end-to-end pipeline guarantees both reproducibility and transparency at every turn.\u003c/p\u003e\u003ch2\u003eGraph API Data Collection\u003c/h2\u003e\u003cp\u003eThe Meta Graph API was used to scrape structured information from Instagram posts such as captions, insights, and hashtags. The initial step was creating an API account using the Meta\u003c/p\u003e\u003cp\u003eDeveloper Platform, where the appropriate permissions were assigned to support API queries. Upon setting these permissions, an access token was created to authenticate API requests.\u003c/p\u003e\u003cp\u003ePython was utilized to retrieve the Instagram account ID, page ID, and other relevant details required to run the API requests. These requests enabled the retrieval of the Instagram account posts and searching via hashtags, thus accumulating all the data needed for the study.\u003c/p\u003e\u003cp\u003eThe data fetching process dynamically pulled useful engagement metrics, including likes, comments, shares, impressions, and reach, depending on the type of media (image, video, or link). Derived insights were embedded in the post data and saved in a well-formatted JSON structure to enable further preprocessing and analysis. This efficient process provided exhaustive data collection for modeling and prediction activities in subsequent phases.\u003c/p\u003e\u003ch2\u003eIntegrating Survey Data\u003c/h2\u003e\u003cp\u003eSurvey data integration is a key implementation step, forming the foundation of user segmentation, behavior analysis, and predictive modeling. The survey collected demographic data (e.g., age, gender, status, device type) and behavior preferences (e.g., content types, video formats, engagement behaviors, social media usage patterns). The data was cleaned, standardized, and structured for consistency, with each element processed separately for analysis with substance.\u003c/p\u003e\u003ch2\u003eAudience Segmentation\u003c/h2\u003e\u003cp\u003eThe data collected in the survey allowed segmentation of users by required demographic factors, including age, gender, and device. Segmentation aimed to compare behavioral patterns across segments of users. The pandas library in Python was employed to compute percentages by segment, generating normalized distributions for each behavioral metric by demographic segment. The findings were then plotted with seaborn heatmaps to identify patterns and inconsistencies between the segments.\u003c/p\u003e\u003ch2\u003eMulti-Variable Pattern Mining\u003c/h2\u003e\u003cp\u003eAlong with single segmentation, multi-variable pattern mining was utilized to find complex patterns among user demographics and content interest. It considered how sets of demographic variables (e.g., gender, age) influence behavior like content consumption and interaction. Through Python, filtered subsets of the data were clustered according to multiple variables, which provided deeper insights into real-world user behavior.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec20\"\u003e\n \u003ch2\u003eAudience Segmentation\u003c/h2\u003e\n \u003cdiv\u003e\n \u003cp\u003eAudience segmentation was performed based on various demographic factors and content prefer- ences. The analysis included the following segments:\u003c/p\u003e\n \u003c/div\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003eFavorite Content Type by Age Group:\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n \u003cdiv\u003e\n \u003cp\u003eThis segment shows how different age groups prefer specific types of content, such as educational, lifestyle, or comedy.\u003c/p\u003e\n \u003c/div\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003eFavorite Content Type by Gender:\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n \u003cdiv\u003e\n \u003cp\u003eThis analysis highlights content type preferences by gender, revealing any significant trends in content consumption.\u003c/p\u003e\n \u003c/div\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003eFavorite Content Type by Status:\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n \u003cdiv\u003e\n \u003cp\u003eHere, the content preferences are segmented by user status (e.g., student, working, other), showing how different groups engage with content.\u003c/p\u003e\n \u003c/div\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003ePreferred Video Length by Age Group: This segment examines how video length preferences vary across different age groups.\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003ePreferred Video Length by Gender:\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n \u003cdiv\u003e\n \u003cp\u003eThis analysis investigates how male and female users differ in their preferred video lengths.\u003c/p\u003e\n \u003c/div\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003ePreferred Video Length by Status:\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n \u003cdiv\u003e\n \u003cp\u003eThis segment looks into how user status (e.g., student, working) influences video length preferences.\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec21\"\u003e\n \u003ch2\u003eMulti-Variable Audience Pattern Mining\u003c/h2\u003e\n \u003cdiv\u003e\n \u003cp\u003eMulti-variable pattern mining was applied to uncover complex relationships between multiple demographic and behavioral variables. The following combinations were analyzed to explore how demographic factors interact with content preferences and engagement behaviors:\u003c/p\u003e\n \u003c/div\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003e \u003cb\u003eAge + Content Type + Preferred Video Length\u003c/b\u003e: This analysis explores how age groups, content preferences, and preferred video lengths intersect to define content engagement patterns.\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003e \u003cb\u003eAge + Content Type + Preferred Time to Use Social Media\u003c/b\u003e: This segment investigates how age, content preferences, and preferred social media usage times correlate to determine\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n \u003cdiv\u003e\n \u003cp\u003epeak engagement periods and content types.\u003c/p\u003e\n \u003c/div\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003e \u003cb\u003eAge + Preferred Time to Use Social Media + Reactions\u003c/b\u003e: This analysis examines how age, preferred usage times, and user reactions (likes, comments, shares) interact, providing insights into when users are most likely to engage with content.\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec22\"\u003e\n \u003ch2\u003eComparison between Graph API Data and Survey Data\u003c/h2\u003e\n \u003cdiv\u003e\n \u003cp\u003eThis section compares the insights derived from the survey data with actual engagement metrics obtained from the **Meta Graph API**. The following comparisons were made to analyze how well survey responses align with real-world engagement behavior:\u003c/p\u003e\n \u003c/div\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003e \u003cb\u003eComparison Between Preferred Usage Time and Actual Engagement\u003c/b\u003e: This analysis examines the alignment between users’ preferred times to use social media (as reported in\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n \u003cdiv\u003e\n \u003cp\u003ethe survey) and the actual engagement metrics (likes, comments, shares) on Instagram posts during those times.\u003c/p\u003e\n \u003c/div\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003e \u003cb\u003eComparison of Preferred Content Type vs. Performance\u003c/b\u003e: This segment compares users’ self-reported favorite content types (e.g., educational, entertainment) with the actual engage- ment metrics (likes, comments, shares) for posts categorized under those content types.\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003e \u003cb\u003eComparison Between Self-Reported and Actual Reactions\u003c/b\u003e: This analysis compares the self-reported reactions (likes, comments, shares) by users in the survey with the actual reactions observed on Instagram posts, assessing how accurately survey responses reflect real-world behavior.\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n\u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eThe objective of this research was to explore the ways in which creators can leverage data-driven strategies on social media sites by examining engagement patterns, audience segmentation, and content effectiveness prediction. By integrating large-scale survey data on user interests with real- time behavioral signals extracted from Instagram (from Meta\u0026rsquo;s Graph API), this study provides an end-to-end framework for informing the types of content that work best for specific demographic segments.\u003c/p\u003e \u003cp\u003eThe findings revealed an evident correlation between users\u0026rsquo; stated content interests and actual Instagram behavior. Educational content emerged as the most favored content category in the survey and simultaneously had the most engagement, particularly when posted during the optimal time frame of 21:00\u0026ndash;00:00. The implication here is the significance of not just developing content aligned with user interests but also publishing it at the right time to achieve optimum visibility and engagement.\u003c/p\u003e \u003cp\u003eThe audience segmentation study, which split users by demographic factors such as age, gender, and status, was very insightful into the way these segments watch content. Younger users (18\u0026ndash;24) liked lifestyle and entertainment content the most strongly, whereas older users (25\u0026ndash;34 and 35+) liked more informative content. These findings imply that content developers can enhance their content strategy by making the tone of their posts appealing to the taste of particular age groups and demographic segments.\u003c/p\u003e \u003cp\u003eMoreover, the study compared survey feedback to Graph API metrics to validate the reliability of self-reported data.\u003c/p\u003e \u003cp\u003eThe high correspondence observed between users\u0026rsquo; stated preferences for particular types of content in the survey and actual use of such content types on Instagram reiterates that survey responses represent a valuable resource for content creators. The high agreement indicates that creators can successfully utilize survey responses to forecast content efficacy and make highly informed decisions regarding posting strategies.\u003c/p\u003e \u003cp\u003eAdditionally, the examination of trends in engagement over time, as illustrated through the graphs, underscored the value of posting content at the times of most intense user activity. Evening posts, specifically those between 9 PM and midnight, received high levels of engagement consis- tently, and this underlined the need for content publishers to time their postings to optimize viewing during these times of peak activity.\u003c/p\u003e \u003cp\u003eIn total, this study offers practical recommendations for how content creators can leverage AI and data analytics to not only know what type of content performs best, but when and where to post to maximize audience interactions.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAppel, G., Grewal, L., Hadi, R., \u0026amp; Stephen, A. T. (2021). The future of social media in marketing. \u003cem\u003eJournal of the Academy of Marketing Science, 48\u003c/em\u003e(1), 79\u0026ndash;95.\u003c/li\u003e\n\u003cli\u003eBaumann, F., Halpern, D., Procaccia, A. D., Rahwan, I., Shapira, I., \u0026amp; Wuthrich, M. (2023). Optimal engagement\u0026ndash;diversity tradeoffs in social media [Preprint]. \u003cem\u003earXiv\u003c/em\u003e. https://doi.org/10.48550/arXiv.2303.03549\u003c/li\u003e\n\u003cli\u003eBerman, T. (2020). AI tools in content creation: Automating engagement and editing. \u003cem\u003eJournal of Content Technology, 10\u003c/em\u003e(3), 220\u0026ndash;235.\u003c/li\u003e\n\u003cli\u003eChen, L., Li, X., Zhang, Y., \u0026amp; Wu, S. (2023). Predicting content success: AI in social media marketing. \u003cem\u003eSocial Media Research Journal, 11\u003c/em\u003e(2), 123\u0026ndash;138.\u003c/li\u003e\n\u003cli\u003eIvanov, D., Volkov, S., Petrov, A., Zhao, D., et al. (2024). Real-time text classification for social media content analysis [Preprint]. https://doi.org/10.13140/RG.2.2.18594.03529\u003c/li\u003e\n\u003cli\u003eJohnson, M., \u0026amp; Patel, R. (2021). The impact of social media on marketing campaigns. \u003cem\u003eInternational \u003c/em\u003e\u003cem\u003eJournal of Marketing, 32\u003c/em\u003e(4), 103\u0026ndash;115.\u003c/li\u003e\n\u003cli\u003eKapoor, K., Tamilmani, K., Rana, N. P., Patil, P. P., Dwivedi, Y. K., \u0026amp; Nerur, S. (2018). Advances in social media research: Past, present and future. \u003cem\u003eInformation Systems Frontiers, 20\u003c/em\u003e(4), 531\u0026ndash;558. https://doi.org/10.1007/s10796-017-9810-y\u003c/li\u003e\n\u003cli\u003eKim, Y. (2021). Predictive models for content performance on social media. \u003cem\u003eInternational Journal of Predictive Analytics, 7\u003c/em\u003e(5), 34\u0026ndash;42.\u003c/li\u003e\n\u003cli\u003eKumar, A., Singh, P., \u0026amp; Verma, R. (2022). Using BERT for text classification in social media content. \u003cem\u003eJournal of AI and Marketing, 8\u003c/em\u003e(3), 56\u0026ndash;68.\u003c/li\u003e\n\u003cli\u003eLang, D. (2017). Using AI for audience segmentation in social media. \u003cem\u003eMedia Analytics, 14\u003c/em\u003e(4), 50\u0026ndash;60.\u003c/li\u003e\n\u003cli\u003eLee, K. (2022). Engagement on Instagram: The role of content format and timing. \u003cem\u003eJournal of Digital Media, 13\u003c/em\u003e(1), 78\u0026ndash;92.\u003c/li\u003e\n\u003cli\u003eSethupathi, G., Akshay, U., Surya, C., \u0026amp; Nath, G. (2019). Integrating social media marketing with AI through predictive examination. \u003cem\u003eInternational Journal of Engineering and Advanced Technology, 9\u003c/em\u003e(1s), 13\u0026ndash;19. https://doi.org/10.35940/ijeat.A1003.1091S19\u003c/li\u003e\n\u003cli\u003eSmith, J. (2020). The rise of social media and content creation. \u003cem\u003eJournal of Media Studies, 5\u003c/em\u003e(2), 45\u0026ndash;60.\u003c/li\u003e\n\u003cli\u003eTran, A., \u0026amp; Nguyen, P. (2021). Machine learning for audience segmentation in social media. \u003cem\u003eJournal of Digital Marketing, 9\u003c/em\u003e(1), 45\u0026ndash;53.\u003c/li\u003e\n\u003cli\u003eWang, S., Menon, S., Long, T., Henderson, K., Li, D., Crowston, K., Hansen, M., Nickerson, J. V., \u0026amp; Chilton, L. B. (2023). ReelReframer: Human\u0026ndash;AI co-creation for news-to-video translation [Preprint]. \u003cem\u003earXiv\u003c/em\u003e. https://doi.org/10.48550/arXiv.2304.09653\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"audience engagement, social media strategy, content optimization, instagram analytics, audience segmentation, social media behavior, meta graph api","lastPublishedDoi":"10.21203/rs.3.rs-6863597/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6863597/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study investigates the relationship between content creation and audience behavior on Instagram, one of the most widely used social media platforms. By combining data from Instagram\u0026rsquo;s Graph API with a user survey, the research examines how individuals from various demographic backgrounds engage with content and identifies key preferences that can guide content creators and business owners in optimizing their strategies.\u003c/p\u003e \u003cp\u003eThe data collected through the Graph API provided insights into the performance of existing posts, while the survey gathered information about user demographics and content preferences. The analysis enabled the identification of optimal posting times, preferred content types, and other factors that influence user engagement.\u003c/p\u003e \u003cp\u003eWhile the study successfully achieved its objectives and offered actionable recommenda- tions for content optimization, there remains potential for further refinement. Future research could expand these findings by incorporating additional data and exploring alternative method- ologies.\u003c/p\u003e","manuscriptTitle":"Optimizing Instagram Engagement: Insights into Content and Audience Behavior","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-11 12:07:44","doi":"10.21203/rs.3.rs-6863597/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"45ef3260-197c-415c-9488-8c26637cd1b3","owner":[],"postedDate":"June 11th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":49822962,"name":"Marketing"}],"tags":[],"updatedAt":"2025-06-11T12:07:44+00:00","versionOfRecord":[],"versionCreatedAt":"2025-06-11 12:07:44","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6863597","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6863597","identity":"rs-6863597","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}