Optimization of Bus Stop Location based on Spatial Analysis and K-Means Clustering Method – A Case Study in Hohhot City

Optimization of Bus Stop Location based on Spatial Analysis and K-Means Clustering Method – A Case Study in Hohhot City | 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 Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Optimization of Bus Stop Location based on Spatial Analysis and K-Means Clustering Method – A Case Study in Hohhot City Jingqiao Yu, Yuan Zhu, Chaowen Wu, Guoqing Song This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3913979/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 The location of bus stops is crucial for attracting bus passengers. Researchers have explored various methods, such as questionnaire surveys and field investigations, to determine the location of bus stops, but these methods largely rely on experience and intuition. With the introduction of big data, a more scientific approach has emerged for siting bus stops. Based on the location data from cellular base stations in Hohhot, China, this paper uses K-means clustering and spatial analysis to calculate the coordinates of bus stops within 4 kilometers of the central business district (CBD) area. Then the study utilizes the city point of interest (POI) data to compare the new bus stops and existing ones in terms of the POI number, bus stop spacing, and POI average distance. The study concludes that the new bus stops outperform the existing ones in terms of POI number and stops spacing. The Analytic Hierarchy Process (AHP) is applied to evaluate the results that indicate the newly constructed bus stops are superior to the existing ones, with an overall optimization rate of 3.194%.The findings of this study can serve as decision-making references for urban planning departments and public transportation operators, aiming to increase public transport passenger flow and improve the traffic conditions in cities. Additionally, the methods utilized in this research can be applied to other cities, assisting them in site selection of bus stops and optimal planning, thereby promoting the development of public transportation. Bus Stops POI Clustering K-means Spatial Analysis Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction As more cities undergo modernization, the number of car ownership of urban residents continues to increase, which has led to traffic congestions of urban roads, caused additional noise and pollution, and yielded more energy consumption (Xue and Yao 2022 ). Public transport is an important way to alleviate these issues, however, modern bus transit systems are facing the challenges such as low passenger willingness to use public transport and low passenger occupancy rates. Typically, residents' willingness to take public transportation is affected by the location of bus stops, the setting of routes, the frequency of bus departures, the comfort of buses, and the price of taking the bus (Ling et al. 2022 ). This paper mainly focuses on the location of the bus stops. Many aspects should be considered in selection of bus stop location. Traditionally, decision makers are willing to divide the areas in the city into several major types, such as parks, hospitals, schools, and scenic spots, in which bus stops are set up in well-known places in each category, such as the main entrance of large shopping malls, school entrances, scenic spots, and community entrances. However, such method relies on experience and intuition to get urban hot spots, rather than objectively proving the willingness of passengers to take the bus. In recent years, the real-time positioning of urban residents can be known through the location data from cellular base stations. The passenger travel hotspots obtained based on this data are proven to be more reliable. Knowing the travel hotspots of passengers can help the decision makers in selections of bus stop location. Literature Reviews In recent years, with the emergence of electric buses and demand-responsive transit service, literature on the optimization of bus stop location has tended to focus on the siting of bus charging stations and the optimization of demand-responsive transit stops (Ahmad et al. 2022 ; Danese et al. 2022 ; P. Li et al. 2022 ; Perumal et al. 2022 ; Shen et al. 2019 ). In addition, research on the siting of integrated transportation hubs has gained more attention compared to traditional bus stop siting studies. However, both electric buses and demand-responsive buses require traditional bus stops for boarding and alighting (Kchaou 2021), traditional bus stops are still necessary for the foreseeable future, making the location of bus stops crucial for the attracting passenger flows to the system. A few recent studies focus on the utilization of big data or city point of interest (POI) data for the optimization of bus stop location. Ji-zhuo et al. ( 2022 )proposed a visualization analysis system called VisB4B, which utilized the fuzzy C-means clustering method to process GPS data and POI data from shared bicycles, generating a heat map of population distribution. The system aimed to consider alternative locations for transit stations; however, specific locations for new bus stops were not determined, and no validation of the results was conducted. Wei and Tao ( 2020 ) utilized mobile phone GPS data and Baidu route planning to calculate passenger origin-destination (OD) linkages and adjusted the positions of bus stops based on the overlap between bus routes and OD linkages. However, the OD results of this study have not been validated, and there was no evidence to demonstrate improved service efficiency of the newly adjusted bus stops. Zu-peng et al. ( 2019 )designed a method to fig out the influence of bus stop location on bus flow trajectory and bus delay under different intersection layouts. However, this research only considered the impact of the intersection layout on the service efficiency of bus stops, and did not consider urban hotspots. Ahmad et al. ( 2022 )designed a hierarchical evaluation method that uses the distance between bus stops as the independent variable and the service level as the dependent variable. However, the assessment of bus stop service levels relied solely on the subjective perception of 120 volunteer participants of similar age. The variables involved were not sufficiently independent, potentially introducing bias into the conclusions. Guangyuan ( 2018 )imported the administrative division, population and POI data into ArcGIS software to display the capacity of bus stops individually or superimposed, so as to obtain the utilization efficiency of bus stops, but did not propose adjustment measures for bus stops with low efficiency. Siji ( 2018 )used the intelligent bus data to analyze the stops queuing probability with statistical methods and made a graph for analysis, introduced the concept of queuing overflow, and proposed a reference for the location setting of high-frequency road bus stops based on the queuing probability. To sum up, existing literature on optimizing bus stop locations is either insufficient in terms of in-depth research and lacks specific optimization results for bus stop selection, or does not compare and evaluate the level of optimization with the original plans. In the paper which devoted to the quantitative evaluation of the service level of the bus stops, Yiyi et al. ( 2022 )calculated the rationality of the overall bus network layout using the entropy-weighted analysis method and validated the results using POI data. Xiaoxiong et al. ( 2019 )developed an application platform that automatically integrates POIs and bus network structural data for visualization analysis. Both articles proposed methods for evaluating bus line networks using POI data, but did not provide specific optimization measures for bus stop selection. Based on the abovementioned studies, this paper proposes a specific new stops selection method for the real bus stops in Hohhot, and quantitatively compares the results. The approach uses clustering and spatial analysis methods and considers multiple factors. There are three innovations of this paper. Firstly, it proposes a new spatial analysis method, the weighted path distance calculation method(WPDM), so that the density of cluster centers is also considered. Secondly, the stops selection of bus stops comprehensively considers multiple influencing factors rather than just a single factor, especially considering the influence of urban hotspots. Lastly, the service evaluation of old and new stops finally verifies the value of this research. DATA The study area selected for this research is the central region of Hohhot, which spans a diameter of 4 kilometers and is characterized by a moderate number of bus stops. The database utilized in this study comprises 460,000 records of cellular base stations, capturing information from hundreds of thousands of users in Hohhot city. The data includes user IDs, timestamps, as well as latitude and longitude coordinates. Sample data is presented in Table 1 . Table 1 Data sample id time longitude latitude 13009506efdf13ac011900b 201807020810 40.818333 111.696111 130095142cf5a670d315ce4 201807020810 40.843888 111.739166 130150099311bf5564e95dd 201807020810 40.867777 111.644166 1301501a71bd769b385a276 201807020810 40.807487 111.654991 1301510fcec2a04e8dbb0cf 201807020810 40.827488 111.685898 The data consists of the location information for residents in Hohhot City, Inner Mongolia, collected in July 2018. To ensure representative data and reduce computational burden, the dataset is divided by day, and only the peak hour data on working days is selected for analysis. Since the base stations’ positioning frequency is not fixed, some residents may have multiple recordings within a day, even if they are stationary at home. Moreover, due to the uneven distribution of base stations, the same resident may be simultaneously recorded by two different base stations. To address these issues, The following methods are employed: 1). For each resident, their location was recorded only once in the same area within a day. 2). A resident’s location was associated with only one base station at any given time. Furthermore, to eliminate coordinate deviation, the data’s coordinate system was converted from GCJ02 to WGS84. After completing the data processing, a total of 453,533 positioning information records were obtained, as illustrated in Fig. 1(a) . Figure 1 Coordinates of Hohhot residents and three level candidate stops The research article is based on the actual road network in Hohhot, which comprises more than 14 roads within its scope. Some of these roads include: Xinhua Street, Zhongshan Road, and others. The study focuses on the real bus stops located along these 14 roads, making them the selected research objects. Methodology The methodology section is presented in the order of computation, including the processes of cluster analysis, the processes of spatial analysis, optimization of bus stop location, and the evaluation of bus stops by Analytic Hierarchy Process(AHP). The Processes of Cluster Analysis Firstly, to facilitate distance calculations between coordinates, it is essential to convert the coordinates from WGS84 to UTM49. In this paper, the K-means clustering method is chosen as it allows control over the number of cluster centers. On one hand, since the data represents the coordinates of base stations, the value of K can be set as close as possible to, but not exceeding, the number of base stations (n = 265). On the other hand, Elbow Method (Syakur et al. 2017 ) is usually used to determine the appropriate K value( Formula 1 ). $$\begin{array}{c}SSE= \sum _{i=1}^{k}\sum _{p\in {C}_{i}}|p-{m}_{i}{|}^{2}\#\left(1\right)\end{array}$$ where SSE is the sum of the squared differences between the distances between the center of each cluster and the sample stops in the cluster. \({C}_{i}\) is the \(i\_th\) cluster, \(p\) is all sample stops in \({C}_{i}\) cluster, and \({m}_{i}\) is the center of the \({C}_{i}\) cluster (the mean of all samples in \({C}_{i}\) ). The determination of K in this paper takes these two aspects into consideration. It is calculated in the interval close to n to obtain the elbow diagram, as shown in Fig. 2 . The abscissa in the figure represents the value of K, and the ordinate is the sum of the squared differences between the distances between the center of each cluster and the sample stops in the cluster under different K values, which is recorded as the value B. When the value of K reaches 263, the curve reaches an inflection point. This implies that increasing K beyond 263 will not significantly enhance the representativeness of the clustering centers for all bus stops. Moreover, the difference between K = 263 and K = 265 is minimal, which justifies the selection of K = 263. Based on this analysis, the coordinate data is subjected to K-means clustering, resulting in 263 clustering centers. Additionally, the weight of each clustering center is recorded based on its density. After clustering, to facilitate the calculation of path distance between coordinates, the coordinate system of the clustering centers is converted back from UTM49 to WGS84 for spatial analysis. The Processes of Spatial Analysis Spatial analysis examines the spatial position, distances, and relationships between objects. In this study, distance analysis of spatial objects is employed to calculate the distances between cluster centers and the candidate bus stops on roads. It is essential to calculate the path distance rather than the Euclidean distance or Manhattan distance. Subsequently, the candidate stops are filtered based on the clustering centers. Therefore, apart from considering the proximity of the true path distance, the weights of the clustering centers also need to be taken into account. The WPDM serves as the core method used in this article for filtering candidate stops. Each cluster center possesses two properties: coordinates and weights. The approach involves computing path distances between each cluster center and every candidate bus stop, identifying the top three candidate stops with the shortest distances from the cluster center, and assigning different weights (e.g., 1, 0.5, and 0.2) based on their ranking (with the candidate stop having the shortest distance assigned a weight of 1, and so forth). Multiple calculations reveal that selecting evenly distributed weights results in insignificant score differences between the selected candidate stops that meet the qualification criteria and the eliminated stops near the qualification threshold. Conversely, choosing more disparate weights concentrates scores too heavily on a few optimal candidate stops, resulting in almost equally low scores for candidate stops above and below the qualification threshold. Hence, the selection of these three weight values (1, 0.5, 0.2) is grounded in empirical findings obtained through repeated calculations. It is important to note that each candidate stop may be selected by multiple cluster centers or not selected by any cluster centers. The stops selected by the cluster center of weight 3 and the cluster center of weight 30 should not be equally important, so the weight of cluster centers should also be assigned to candidate stops. The specific method is to multiply the weight of each candidate stop above by the weight of its corresponding cluster center to obtain a new weight named \({W}_{j,S}\) (Formula 2) . $$\begin{array}{c}{W}_{j,S}= \sum _{i=1}^{k}\sum _{j=1}^{Q}{\omega }_{i}\times {\omega }_{j,S}\#（2）\end{array}$$ where \({\omega }_{i}\) represents the weight of the cluster center, and its value is the density of the cluster center. k is the number of cluster centers. \({\omega }_{j,s}\) represents the weight of candidate stop j in stage S. For example, \({\omega }_{\text{87,1}}\) represents the weight of the candidate stop 87 in first-level, The value of S is 1,2. k is the number of cluster centers. Q is the number of candidate stops. \({W}_{j,S}\) represents the total weight of the candidate stops in S-level. The next task is to rank each candidate stop in descending order based on the \({W}_{j,S}\) values (Table 2 ). This ranking reflects the combined influence of two factors: the frequency of candidate stop selection by clustering centers and the weights of the clustering centers making those selection. In essence, it signifies the importance of each candidate stop in the context of the analysis. Table 2 \({\varvec{W}}_{\varvec{j}}\) for candidate stops Candidate stop ID longitude latitude \({W}_{j,S}\) order 47 111.6603509 40.81260974 19.63 1 29 111.6716194 40.8218115 15.717 2 149 111.6639795 40.81414413 14.892 3 85 111.6480058 40.80694325 14.077 4 37 111.6652291 40.81660316 13.989 5 Optimization of Bus Stop’s Location In this section, we employ the WPDM to compute and filter candidate bus stops. On the 14 roads in Hohhot, a candidate stop is selected every 100m, therefore a total of 328 first-level candidate stops are obtained as shown in Fig. 1(b) . According to WPDM, the weighted path distances are calculated between 328 primary candidate stops and 263 cluster centers. The resulting values of total weight \({W}_{j,1}\) are obtained. The next step involves sorting the candidate stops in descending order based on the values of \({W}_{j,1}\) , and retaining the top 100 stops with the highest \({W}_{j,1}\) values, while eliminating the other stops. These 100 stops are then expanded within a range of 50 meters before and after the road, resulting in 234 second-level candidate stops that are spaced 50 meters apart, as shown in Fig. 1(c) . Likewise, WPDM is used to calculate between the second-level candidate stops and cluster centers. In this screening process, only the two candidate stops closest to each cluster center are retained, and they are assigned weights of 1 and 0.2 respectively. Afterward, the total weights \({W}_{j,2}\) of the stops are calculated, and the top 100 stops with the highest \({W}_{j,2}\) are selected and named as third-level candidate stops as shown in Fig. 1(d) . Figure 3 shows the entire filtering and calculation process from first-level candidate stops to third-level candidate stops. To eliminate adjacent stops that are too close in distance, the level three candidate stops are divided and labeled according to the routes. The path distance between adjacent candidate stops on each route is calculated. Since Hohhot's city center has a large passenger flow, a spacing of 100 meters between stops is considered an acceptable minimum distance. Therefore, adjacent candidate stops with less than a 100m interval are compared, and the stop with a higher weight is retained while the stop with a lower weight is eliminated. After filtering, the stops that meet the criteria are selected by the value of \({W}_{j,2}\) , retaining those with higher \({W}_{j,2}\) and eliminating others, resulting in the desired number of new bus stops. These stops represent the newly obtained bus stops in this study, which is equal to the number of actual bus stops within the scope of the research in Hohhot. The Evaluation of Bus Stops by AHP This section describes the evaluation method for both the original bus stops in Hohhot City and the newly obtained bus stops in this study. Three indicators are used for evaluating the bus stops in this study: the number of POIs within the service range of a bus stop(POI number), the distance between adjacent bus stops(stops distances), and the average distance between the stop and POIs within the stop's service area (stop-POIs distances). The weights for these three indicators are calculated using the Analytic Hierarchy Process (AHP). This method combines quantitative analysis with qualitative analysis by incorporating the decision maker's expertise to determine the relative importance of each evaluation criterion and assign appropriate weightings (J. Li and Zou 2011 ). The results are presented in Table 3 . Table 3 The weight of the three indicators Indicator POI number stops distances stop-POI distances Weight (C i ) 0.63334572 0.260497956 0.106156324 A total of 1769 POI records are used for evaluation. The POI dataset includes the POI's ID, type, and coordinates. The types of POIs include schools, shopping malls, parks, companies, government institutions, etc. The distribution of numbers of POIs within a 1km radius of bus stops follows an exponential distribution with R-square 0.9914. Bus stops located in the city center have a higher number of POIs, but a lower average clustering coefficient (Zhang et al. 2021 ), indicating a more evenly distributed POI density in the city center. Therefore, it is necessary to assign different weights to bus stops in order to differentiate the importance of POI quantities in different distance categories. Preliminary calculations reveal that within the study area, there are no more than 60 POIs within a 200m radius of bus stops in Hohhot, and there are very few bus stops with over 30 POIs in their service range. Based on this, for the indicator of POI number, the following scoring criteria are established: a score of 0.3 is assigned if there are 0–10 POIs within the range, a score of 0.6 is assigned if there are 11–20 POIs, and a score of 1 is assigned if there are more than 30 POIs. Since both criteria consist of 64 bus stops, the full score for the 64 bus stops should be 64 points. To convert this proportionally to 10 points, the scoring standards are presented in Table 4 . Table 4 The assignment of POI number POI number range Assign points 0–10 0.3 11–20 0.6 21–30 0.8 > 30 1 Full marks 64 Standard full marks 10 The spacing between bus stops affects the operating speed and service efficiency of buses. As the focus of this study is the city center of Hohhot, an area with inadequate and congested transportation infrastructure, bus speeds are not particularly fast. Therefore, the appropriate distance between adjacent bus stops is determined to be between 100 and 500m (Tirachini 2014 ). The distances between neighboring bus stops on each route are calculated, and scoring is assigned to the bus stops based on distance categories. The specific scores are subjectively determined by the author's years of experience living in Hohhot. The specific scoring criteria are presented in Table 5 . Table 5 The assignment of bus spacing bus spacing range Assign points 0-100m 0.5 101-500m 1 501-1000m 0.5 > 1000m 0.2 Full marks 64 Standard full marks 10 For the stop-POIs distances indicator, the average distance between each bus stop and all POIs within a 200m radius of the stop is calculated. When this average distance is not more than 100 meters, it indicates that the bus stop is located near several major POIs, which is the ideal situation, earning a score of 1. In all other cases, the attractiveness of the bus stop to passengers slightly decreases. The scores are assigned based on the author's experience living in Hohhot, as shown in Table 6 . Table 6 The assignment of stops-POI distances stop-POI distances range Assign points 150 0.5 Full marks 64 Standard full marks 10 The scores of the original and new bus stops are separately calculated and normalized in three aspects, ensuring that the scores are distributed within the range of [0, 10]. Then, based on the weights obtained from the AHP (Table 3 ), the overall evaluation score is calculated using Formula 3 . $$\begin{array}{c}P= \sum _{l=1}^{3}{C}_{l}\times {S}_{l}\#\left(3\right)\end{array}$$ P represents the total score, \({C}_{l}\) represents the weights of the three indicators, which respectively indicate the importance of the three indicators, and \({S}_{l}\) represents the scores of the old or new bus stops within the three indicators. RESULTS The real 64 bus stops in Hohhot(original stops) are compared with the 64 new bus stops based on the three indicators mentioned above(Fig. 4 ). In terms of POI numbers, the new stops have an advantage over the original stops. The total number of POIs within the service range of the original bus stops is 935, while the total number of POIs within the service range of the new bus stops is 965. Therefore, the new bus stops serve 30 more POIs than the original bus stops. On average, each original bus stop serves 14.609 POIs, while the average number of POIs served by each new bus stop is 15.078. This indicates that each new bus stop serves an additional 0.5 POIs compared to the original bus stops. After scoring, using 10 as the maximum score, the original bus stops achieve a score of 5.406, while the new bus stops achieve a score of 5.594 (Table 7 ). Table 7 Comparison of old and new stops POI number sum average scores( \({S}_{1}\) , \({S}_{1}^{{\prime }}\) ) original stops 935 14.609 5.406 new stops 965 15.078 5.594 bus stops spacing median(m) average(m) scores( \({S}_{2}\) , \({S}_{2}^{{\prime }}\) ) original stops 376.7334646 426.2 6.359 new stops 308.8874551 408.1 6.797 POI distance median(m) average(m) scores( \({S}_{3}\) , \({S}_{3}^{{\prime }}\) ) original stops 122.245 122.536 7.328 new stops 131.804 130.194 6.891 P(final score) original stops 5.8583 new stops 6.0454 In terms of bus stop spacing, the average distance between adjacent original bus stops is 426.20 meters, while the average distance between adjacent new bus stops is 408.10 meters. This indicates that the spacing between new bus stops is smaller. The median and mean values of bus stop spacing for both original and new bus stops fall within the normal range, but the spacing between new bus stops generally aligns better with the typical bus stop spacing in urban city centers. The original bus stops achieve a score of 6.359, while the new bus stops achieve a score of 6.797 (Table 7 ). In terms of POI distance, the new stops have a disadvantage compared to the original stops. The average distance between all POIs within the service range of the original stop and the stop itself is 122.536 meters, while the average distance between all POIs within the service range of the new stop and the stop itself is 130.194 meters, which means that the new stop has an average distance of 7.5 meters longer than the original stop; The median score of the original stop is 122.245 meters, and the median score of the new stop is 131.804 meters, indicating that the new stop does not have an advantage in terms of POI distance. After scoring, with a full score of 10, the original stop scored 7.328, and the new stop scored 6.891 (Table 7 ). Based on the scores of the new and original stops in the above three indicators, the total score can be calculated using the formula 3 . The final total score of bus stops is shown in Table 7 . The comprehensive score of the original stops in the three indicators is 5.853, and for the new stops, the value is 6.0454. The new stops scored higher and was optimized by 3.194%. The conclusions drawn support the rationality of this research method, which suggests that the new stop is better than the original stop in terms of service quality, convenience, and overall user experience. The limited optimization scope may be due to the sparse distribution of POI data within the study area. If the POI data were more densely distributed, the evaluation of the two schemes would likely be more accurate, and the scoring differences between the original and new bus stops would be more pronounced. Conclusions This paper focuses on 14 bus routes and 64 bus stops within a 4-kilometer radius in the downtown area of Hohhot. Multisource data is used, including POI, the spacing distribution of bus stop and mobile signaling data from residents of Hohhot. The study employs the K-means clustering and the WPDM to calculate new bus stop locations. The new bus stops outperform the existing ones in terms of the POIs number and bus stop spacing, resulting in an overall optimization of 3.194% compared to the original bus stops in Hohhot. Due to limitations in the data used in this study, the original location data cannot accurately represent the distribution characteristics of bus passenger flows. Consequently, the optimization level of the bus stop selection method in this study is relatively low. Therefore, finding more optimal data sources will be a key focus for future research. Additionally, expanding the geographic scope of the study would enhance its value, and there is room for improvement in the scoring method for evaluation indicators. Declarations Ethical Approval This declaration is “not applicable”. Competing interests No conflict of interest exits in the submission of this manuscript, and manuscript is approved by all authors for publication. Authors' contributions JY wrote the main manuscript text and prepared all figures. ZY revised the initial draft and modified Table 7. CW and GS collected and organized references. All authors read and approved the final manuscript. Funding Our study is supported by Young Scientists Fund of the National Natural Science Foundation of China (Grant No. 61903205) and Young Scientists Fund of Natural Science of Inner Mongolia (Grant No. 2019BS07002). Data Availability Statement The cell phone mobility data was acquired from China Unicom, and they have not given their permission for researchers to share their data. The POI data was acquired from Gaode Map, data requests can be made to Gaode Map via this website: https://lbs.amap.com/. References Ahmad, F., Iqbal, A., Ashraf, I., & Marzband, M. (2022). Optimal location of electric vehicle charging station and its impact on distribution network: A review. Energy Reports, 8 , 2314-2333. Danese, A., Torsæter, B. N., Sumper, A., & Garau, M. (2022). Planning of high-power charging stations for electric vehicles: A review. Applied Sciences, 12 (7), 3214. Guangyuan, Y. (2018). Evaluation method of current capacity of urban bus stops based on POI data and ArcGIS Spatial analysis technology. Beauty & Times (4), 67-68. Ji-zhuo, L., Ting, X., & Min, Z. (2022). Visualization Analysis of Bus Stop Optimization Based on Multi-source Data. Journal of Chinese Computer Systems, 43 (7). Kchaou-Boujelben, M. (2021). Charging station location problem: A comprehensive review on models and solution approaches. Transportation Research Part C: Emerging Technologies 132 , 103376. Li, J., & Zou, P. X. W. (2011). Fuzzy AHP-Based Risk Assessment Methodology for PPP Projects. Journal of Construction Engineering and Management-Asce, 137 (12), 1205-1209. doi:10.1061/(asce)co.1943-7862.0000362 Li, P., Jiang, L., Zhang, S., & Jiang, X. (2022). Demand Response Transit Scheduling Research Based on Urban and Rural Transportation Station Optimization. Sustainability, 14 (20), 13328. Ling, S., Ma, S., & Jia, N. (2022). Sustainable urban transportation development in China: A behavioral perspective. Frontiers of Engineering Management, 9 (1), 16-30. Perumal, S. S., Lusby, R. M., & Larsen, J. (2022). Electric bus planning & scheduling: A review of related problems and methodologies. European Journal of Operational Research, 301 (2), 395-413. Shen, Z.-J. M., Feng, B., Mao, C., & Ran, L. (2019). Optimization models for electric vehicle service operations: A literature review. Transportation Research Part B: Methodological 128 , 462-477. Siji, B. (2018). Research on bus entry queuing probability and station location based on intelligent bus data. (Master Master), Southwest Jiaotong University, Available from Cnki Syakur, M. A., Khotimah, B. K., Rochman, E. M. S., & Satoto, B. D. (2017, 2018 Nov09). Integration K-Means Clustering Method and Elbow Method For Identification of The Best Customer Profile Cluster. Paper presented at the 2nd International Conference on Vocational Education and Electrical Engineering (ICVEE), Univ Negeri Surabaya, Fac Engn, Dept Elect Engn, Surabaya, INDONESIA. Tirachini, A. (2014). The economics and engineering of bus stops: Spacing, design and congestion. Transportation research part A: policy practice 59 , 37-57. Wei, C., & Tao, L. (2020). Optimization of bus stops based on multi-source Big data. 2020 China Urban Transportation planning Annual Conference , 1070-1080. doi:10.26914/c.cnkihy.2020.032165 Xiaoxiong, W., Yongxin, L., & Jukang, L. (2019). Research on big data mining based evaluation method for life service level of urban public transport stops. Modern Electronics Technique, 42 (02), 67-70. doi:10.16652/j.issn.1004⁃373x.2019.02.016 Xue, F., & Yao, E. (2022). Impact analysis of residential relocation on ownership, usage, and carbon-dioxide emissions of private cars. Energy Reports, 252 , 124110. Yiyi, C., Jianhua, G., & Huanxin, J. (2022). An Evaluation of Bus Station Layout Rationality Using Matter Element Analysis Based on Points of Interest Data. Journal of Transport Information and Safety, 38 (06), 63-72. doi:10. 3963/j. jssn. 1674-4861. 2020. 06. 009 Zhang, H., Li, X., Zhang, L., Wang, W., Jia, J., & Shi, B. (2021). Discovering station patterns of urban transit network with multisource data: empirical evidence in Jinan, China. KSCE Journal of Civil Engineering 25 (2), 680-691. Zu-peng, L., Ke-ping, L., Ya-qin, H., & Cai-xia, X. (2019). Optimization of Bus Stop Layout Based on Time Distance Trajectory. Journal of Highway and Transportation Ｒesearch and Development, 36 (6), 103-111. Additional Declarations No competing interests reported. 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 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-3913979","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":274434567,"identity":"d9cc6fac-46ae-41d4-a50a-b476a885f3e4","order_by":0,"name":"Jingqiao Yu","email":"","orcid":"","institution":"Inner Mongolia University","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Jingqiao","middleName":"","lastName":"Yu","suffix":""},{"id":274434568,"identity":"b22c0a86-167a-470b-b7ac-70e552e527c8","order_by":1,"name":"Yuan Zhu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAxElEQVRIiWNgGAWjYDACCQglxwam2EjQYky6lsQGorXIz+4xk+apuZPeJ3bGgOFD2WEG/tkN+LUwzjkD1HLsWW6bdI4B44xzhxkk7hzAr4VZIgeohe0wWAszb9thBgOJBPxa2MBa/h1OZwNp+UuMFh6QFqDhCWAtjMRokZBIK7ac23fYsE06reBgz7l0HokbBLTIz0jeeOPNt8Py8rOTNz74UWYtxz+DgBYgYJGAsQ6AXEpQPRAwfyBG1SgYBaNgFIxgAAA4PTm6OOyB4QAAAABJRU5ErkJggg==","orcid":"","institution":"Inner Mongolia Center for Transportation Research, Inner Mongolia University","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Yuan","middleName":"","lastName":"Zhu","suffix":""},{"id":274434569,"identity":"5e5a9f42-a2b6-4247-849f-f75cdc6928a6","order_by":2,"name":"Chaowen Wu","email":"","orcid":"","institution":"Inner Mongolia University","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Chaowen","middleName":"","lastName":"Wu","suffix":""},{"id":274434570,"identity":"dd74a557-109a-442b-8daa-611cdd577114","order_by":3,"name":"Guoqing Song","email":"","orcid":"","institution":"Inner Mongolia University","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Guoqing","middleName":"","lastName":"Song","suffix":""}],"badges":[],"createdAt":"2024-01-31 13:34:53","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3913979/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3913979/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":51715836,"identity":"3b967c98-c7f6-4fd9-95fe-e69c7e65285a","added_by":"auto","created_at":"2024-02-27 21:01:07","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":188554,"visible":true,"origin":"","legend":"\u003cp\u003eCoordinates of Hohhot residents and three level candidate stops\u003c/p\u003e","description":"","filename":"F1.png","url":"https://assets-eu.researchsquare.com/files/rs-3913979/v1/d40db91a6cb90bdc0612194a.png"},{"id":51715837,"identity":"98126be4-ca11-4ff0-a43d-2423900b110b","added_by":"auto","created_at":"2024-02-27 21:01:07","extension":"jpeg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":141367,"visible":true,"origin":"","legend":"\u003cp\u003eVariation of clustering efficiency with K value(Elbow diagram)\u003c/p\u003e","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-3913979/v1/a67f7aff75f1c4a4d8d876f9.jpeg"},{"id":51715835,"identity":"2935c4b6-7b1e-4e59-90b7-b88f67f3d58f","added_by":"auto","created_at":"2024-02-27 21:01:07","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":60173,"visible":true,"origin":"","legend":"\u003cp\u003eSelection process of candidate stops\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-3913979/v1/104525ee261a67df8d21c225.png"},{"id":51715838,"identity":"e7cfa946-bc0a-4adb-a46f-d95dc6f96401","added_by":"auto","created_at":"2024-02-27 21:01:07","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":319755,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of bus stops before and after calculation\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-3913979/v1/069774834fe1769973044f62.png"},{"id":53816425,"identity":"a158f90c-f530-4b3a-9975-1f13be711031","added_by":"auto","created_at":"2024-03-31 18:09:19","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":989650,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3913979/v1/2af69a9c-54f3-478c-a832-99e38c822f1f.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Optimization of Bus Stop Location based on Spatial Analysis and K-Means Clustering Method – A Case Study in Hohhot City","fulltext":[{"header":"Introduction","content":"\u003cp\u003eAs more cities undergo modernization, the number of car ownership of urban residents continues to increase, which has led to traffic congestions of urban roads, caused additional noise and pollution, and yielded more energy consumption (Xue and Yao \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Public transport is an important way to alleviate these issues, however, modern bus transit systems are facing the challenges such as low passenger willingness to use public transport and low passenger occupancy rates. Typically, residents' willingness to take public transportation is affected by the location of bus stops, the setting of routes, the frequency of bus departures, the comfort of buses, and the price of taking the bus (Ling et al. \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). This paper mainly focuses on the location of the bus stops.\u003c/p\u003e \u003cp\u003eMany aspects should be considered in selection of bus stop location. Traditionally, decision makers are willing to divide the areas in the city into several major types, such as parks, hospitals, schools, and scenic spots, in which bus stops are set up in well-known places in each category, such as the main entrance of large shopping malls, school entrances, scenic spots, and community entrances. However, such method relies on experience and intuition to get urban hot spots, rather than objectively proving the willingness of passengers to take the bus. In recent years, the real-time positioning of urban residents can be known through the location data from cellular base stations. The passenger travel hotspots obtained based on this data are proven to be more reliable. Knowing the travel hotspots of passengers can help the decision makers in selections of bus stop location.\u003c/p\u003e\n\u003ch3\u003eLiterature Reviews\u003c/h3\u003e\n\u003cp\u003eIn recent years, with the emergence of electric buses and demand-responsive transit service, literature on the optimization of bus stop location has tended to focus on the siting of bus charging stations and the optimization of demand-responsive transit stops (Ahmad et al. \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Danese et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; P. Li et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Perumal et al. \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Shen et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). In addition, research on the siting of integrated transportation hubs has gained more attention compared to traditional bus stop siting studies. However, both electric buses and demand-responsive buses require traditional bus stops for boarding and alighting (Kchaou 2021), traditional bus stops are still necessary for the foreseeable future, making the location of bus stops crucial for the attracting passenger flows to the system.\u003c/p\u003e \u003cp\u003eA few recent studies focus on the utilization of big data or city point of interest (POI) data for the optimization of bus stop location. Ji-zhuo et al. (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2022\u003c/span\u003e)proposed a visualization analysis system called VisB4B, which utilized the fuzzy C-means clustering method to process GPS data and POI data from shared bicycles, generating a heat map of population distribution. The system aimed to consider alternative locations for transit stations; however, specific locations for new bus stops were not determined, and no validation of the results was conducted. Wei and Tao (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) utilized mobile phone GPS data and Baidu route planning to calculate passenger origin-destination (OD) linkages and adjusted the positions of bus stops based on the overlap between bus routes and OD linkages. However, the OD results of this study have not been validated, and there was no evidence to demonstrate improved service efficiency of the newly adjusted bus stops. Zu-peng et al. (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2019\u003c/span\u003e)designed a method to fig out the influence of bus stop location on bus flow trajectory and bus delay under different intersection layouts. However, this research only considered the impact of the intersection layout on the service efficiency of bus stops, and did not consider urban hotspots. Ahmad et al. (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2022\u003c/span\u003e)designed a hierarchical evaluation method that uses the distance between bus stops as the independent variable and the service level as the dependent variable. However, the assessment of bus stop service levels relied solely on the subjective perception of 120 volunteer participants of similar age. The variables involved were not sufficiently independent, potentially introducing bias into the conclusions. Guangyuan (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2018\u003c/span\u003e)imported the administrative division, population and POI data into ArcGIS software to display the capacity of bus stops individually or superimposed, so as to obtain the utilization efficiency of bus stops, but did not propose adjustment measures for bus stops with low efficiency. Siji (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2018\u003c/span\u003e)used the intelligent bus data to analyze the stops queuing probability with statistical methods and made a graph for analysis, introduced the concept of queuing overflow, and proposed a reference for the location setting of high-frequency road bus stops based on the queuing probability. To sum up, existing literature on optimizing bus stop locations is either insufficient in terms of in-depth research and lacks specific optimization results for bus stop selection, or does not compare and evaluate the level of optimization with the original plans.\u003c/p\u003e \u003cp\u003eIn the paper which devoted to the quantitative evaluation of the service level of the bus stops, Yiyi et al. (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2022\u003c/span\u003e)calculated the rationality of the overall bus network layout using the entropy-weighted analysis method and validated the results using POI data. Xiaoxiong et al. (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2019\u003c/span\u003e)developed an application platform that automatically integrates POIs and bus network structural data for visualization analysis. Both articles proposed methods for evaluating bus line networks using POI data, but did not provide specific optimization measures for bus stop selection.\u003c/p\u003e \u003cp\u003eBased on the abovementioned studies, this paper proposes a specific new stops selection method for the real bus stops in Hohhot, and quantitatively compares the results. The approach uses clustering and spatial analysis methods and considers multiple factors. There are three innovations of this paper. Firstly, it proposes a new spatial analysis method, the weighted path distance calculation method(WPDM), so that the density of cluster centers is also considered. Secondly, the stops selection of bus stops comprehensively considers multiple influencing factors rather than just a single factor, especially considering the influence of urban hotspots. Lastly, the service evaluation of old and new stops finally verifies the value of this research.\u003c/p\u003e\n\u003ch3\u003eDATA\u003c/h3\u003e\n\u003cp\u003eThe study area selected for this research is the central region of Hohhot, which spans a diameter of 4 kilometers and is characterized by a moderate number of bus stops. The database utilized in this study comprises 460,000 records of cellular base stations, capturing information from hundreds of thousands of users in Hohhot city. The data includes user IDs, timestamps, as well as latitude and longitude coordinates. Sample data is presented 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\u003eData sample\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=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eid\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003etime\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003elongitude\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003elatitude\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e13009506efdf13ac011900b\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e201807020810\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e40.818333\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e111.696111\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e130095142cf5a670d315ce4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e201807020810\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e40.843888\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e111.739166\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e130150099311bf5564e95dd\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e201807020810\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e40.867777\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e111.644166\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1301501a71bd769b385a276\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e201807020810\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e40.807487\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e111.654991\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1301510fcec2a04e8dbb0cf\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e201807020810\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e40.827488\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e111.685898\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe data consists of the location information for residents in Hohhot City, Inner Mongolia, collected in July 2018. To ensure representative data and reduce computational burden, the dataset is divided by day, and only the peak hour data on working days is selected for analysis. Since the base stations\u0026rsquo; positioning frequency is not fixed, some residents may have multiple recordings within a day, even if they are stationary at home. Moreover, due to the uneven distribution of base stations, the same resident may be simultaneously recorded by two different base stations. To address these issues, The following methods are employed: 1). For each resident, their location was recorded only once in the same area within a day. 2). A resident\u0026rsquo;s location was associated with only one base station at any given time. Furthermore, to eliminate coordinate deviation, the data\u0026rsquo;s coordinate system was converted from GCJ02 to WGS84. After completing the data processing, a total of 453,533 positioning information records were obtained, as illustrated in \u003cb\u003eFig.\u0026nbsp;1(a)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003e \u003cb\u003eFigure\u0026nbsp;1\u003c/b\u003e Coordinates of Hohhot residents and three level candidate stops\u003c/p\u003e \u003cp\u003eThe research article is based on the actual road network in Hohhot, which comprises more than 14 roads within its scope. Some of these roads include: Xinhua Street, Zhongshan Road, and others. The study focuses on the real bus stops located along these 14 roads, making them the selected research objects.\u003c/p\u003e"},{"header":"Methodology","content":"\u003cp\u003eThe \u003cspan refid=\"Sec4\" class=\"InternalRef\"\u003emethodology\u003c/span\u003e section is presented in the order of computation, including the processes of cluster analysis, the processes of spatial analysis, optimization of bus stop location, and the evaluation of bus stops by Analytic Hierarchy Process(AHP).\u003c/p\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eThe Processes of Cluster Analysis\u003c/h2\u003e \u003cp\u003eFirstly, to facilitate distance calculations between coordinates, it is essential to convert the coordinates from WGS84 to UTM49. In this paper, the K-means clustering method is chosen as it allows control over the number of cluster centers. On one hand, since the data represents the coordinates of base stations, the value of K can be set as close as possible to, but not exceeding, the number of base stations (n\u0026thinsp;=\u0026thinsp;265). On the other hand, Elbow Method (Syakur et al. \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2017\u003c/span\u003e) is usually used to determine the appropriate K value(\u003cb\u003eFormula 1\u003c/b\u003e).\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\begin{array}{c}SSE= \\sum _{i=1}^{k}\\sum _{p\\in {C}_{i}}|p-{m}_{i}{|}^{2}\\#\\left(1\\right)\\end{array}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere SSE is the sum of the squared differences between the distances between the center of each cluster and the sample stops in the cluster. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({C}_{i}\\)\u003c/span\u003e\u003c/span\u003e is the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(i\\_th\\)\u003c/span\u003e\u003c/span\u003e cluster, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(p\\)\u003c/span\u003e\u003c/span\u003e is all sample stops in \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({C}_{i}\\)\u003c/span\u003e\u003c/span\u003e cluster, and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({m}_{i}\\)\u003c/span\u003e\u003c/span\u003e is the center of the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({C}_{i}\\)\u003c/span\u003e\u003c/span\u003e cluster (the mean of all samples in \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({C}_{i}\\)\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe determination of K in this paper takes these two aspects into consideration. It is calculated in the interval close to n to obtain the elbow diagram, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003e. The abscissa in the figure represents the value of K, and the ordinate is the sum of the squared differences between the distances between the center of each cluster and the sample stops in the cluster under different K values, which is recorded as the value B.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWhen the value of K reaches 263, the curve reaches an inflection point. This implies that increasing K beyond 263 will not significantly enhance the representativeness of the clustering centers for all bus stops. Moreover, the difference between K\u0026thinsp;=\u0026thinsp;263 and K\u0026thinsp;=\u0026thinsp;265 is minimal, which justifies the selection of K\u0026thinsp;=\u0026thinsp;263.\u003c/p\u003e \u003cp\u003eBased on this analysis, the coordinate data is subjected to K-means clustering, resulting in 263 clustering centers. Additionally, the weight of each clustering center is recorded based on its density. After clustering, to facilitate the calculation of path distance between coordinates, the coordinate system of the clustering centers is converted back from UTM49 to WGS84 for spatial analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eThe Processes of Spatial Analysis\u003c/h2\u003e \u003cp\u003eSpatial analysis examines the spatial position, distances, and relationships between objects. In this study, distance analysis of spatial objects is employed to calculate the distances between cluster centers and the candidate bus stops on roads. It is essential to calculate the path distance rather than the Euclidean distance or Manhattan distance.\u003c/p\u003e \u003cp\u003eSubsequently, the candidate stops are filtered based on the clustering centers. Therefore, apart from considering the proximity of the true path distance, the weights of the clustering centers also need to be taken into account. The WPDM serves as the core method used in this article for filtering candidate stops.\u003c/p\u003e \u003cp\u003eEach cluster center possesses two properties: coordinates and weights. The approach involves computing path distances between each cluster center and every candidate bus stop, identifying the top three candidate stops with the shortest distances from the cluster center, and assigning different weights (e.g., 1, 0.5, and 0.2) based on their ranking (with the candidate stop having the shortest distance assigned a weight of 1, and so forth). Multiple calculations reveal that selecting evenly distributed weights results in insignificant score differences between the selected candidate stops that meet the qualification criteria and the eliminated stops near the qualification threshold. Conversely, choosing more disparate weights concentrates scores too heavily on a few optimal candidate stops, resulting in almost equally low scores for candidate stops above and below the qualification threshold. Hence, the selection of these three weight values (1, 0.5, 0.2) is grounded in empirical findings obtained through repeated calculations. It is important to note that each candidate stop may be selected by multiple cluster centers or not selected by any cluster centers.\u003c/p\u003e \u003cp\u003eThe stops selected by the cluster center of weight 3 and the cluster center of weight 30 should not be equally important, so the weight of cluster centers should also be assigned to candidate stops. The specific method is to multiply the weight of each candidate stop above by the weight of its corresponding cluster center to obtain a new weight named \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({W}_{j,S}\\)\u003c/span\u003e\u003c/span\u003e\u003cb\u003e(Formula 2)\u003c/b\u003e .\u003cdiv id=\"Equb\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equb\" name=\"EquationSource\"\u003e\n$$\\begin{array}{c}{W}_{j,S}= \\sum _{i=1}^{k}\\sum _{j=1}^{Q}{\\omega }_{i}\\times {\\omega }_{j,S}\\#（2）\\end{array}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({\\omega }_{i}\\)\u003c/span\u003e\u003c/span\u003e represents the weight of the cluster center, and its value is the density of the cluster center. k is the number of cluster centers.\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({\\omega }_{j,s}\\)\u003c/span\u003e\u003c/span\u003e represents the weight of candidate stop j in stage S. For example, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({\\omega }_{\\text{87,1}}\\)\u003c/span\u003e\u003c/span\u003e represents the weight of the candidate stop 87 in first-level, The value of S is 1,2. k is the number of cluster centers. Q is the number of candidate stops. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({W}_{j,S}\\)\u003c/span\u003e\u003c/span\u003e represents the total weight of the candidate stops in S-level.\u003c/p\u003e \u003cp\u003eThe next task is to rank each candidate stop in descending order based on the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({W}_{j,S}\\)\u003c/span\u003e\u003c/span\u003e values (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). This ranking reflects the combined influence of two factors: the frequency of candidate stop selection by clustering centers and the weights of the clustering centers making those selection. In essence, it signifies the importance of each candidate stop in the context of the analysis.\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\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({\\varvec{W}}_{\\varvec{j}}\\)\u003c/span\u003e\u003c/span\u003e for candidate stops\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=\"char\" char=\".\" 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=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCandidate stop ID\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003elongitude\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003elatitude\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({W}_{j,S}\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eorder\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e111.6603509\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e40.81260974\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e19.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e111.6716194\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e40.8218115\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e15.717\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e149\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e111.6639795\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e40.81414413\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e14.892\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e111.6480058\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e40.80694325\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e14.077\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e111.6652291\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e40.81660316\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e13.989\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e5\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=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eOptimization of Bus Stop\u0026rsquo;s Location\u003c/h2\u003e \u003cp\u003eIn this section, we employ the WPDM to compute and filter candidate bus stops. On the 14 roads in Hohhot, a candidate stop is selected every 100m, therefore a total of 328 first-level candidate stops are obtained as shown in \u003cb\u003eFig.\u0026nbsp;1(b)\u003c/b\u003e.\u003c/p\u003e \u003cp\u003eAccording to WPDM, the weighted path distances are calculated between 328 primary candidate stops and 263 cluster centers. The resulting values of total weight \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({W}_{j,1}\\)\u003c/span\u003e\u003c/span\u003eare obtained. The next step involves sorting the candidate stops in descending order based on the values of\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({W}_{j,1}\\)\u003c/span\u003e\u003c/span\u003e, and retaining the top 100 stops with the highest \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({W}_{j,1}\\)\u003c/span\u003e\u003c/span\u003evalues, while eliminating the other stops. These 100 stops are then expanded within a range of 50 meters before and after the road, resulting in 234 second-level candidate stops that are spaced 50 meters apart, as shown in \u003cb\u003eFig.\u0026nbsp;1(c)\u003c/b\u003e.\u003c/p\u003e \u003cp\u003eLikewise, WPDM is used to calculate between the second-level candidate stops and cluster centers. In this screening process, only the two candidate stops closest to each cluster center are retained, and they are assigned weights of 1 and 0.2 respectively. Afterward, the total weights \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({W}_{j,2}\\)\u003c/span\u003e\u003c/span\u003e of the stops are calculated, and the top 100 stops with the highest \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({W}_{j,2}\\)\u003c/span\u003e\u003c/span\u003e are selected and named as third-level candidate stops as shown in \u003cb\u003eFig.\u0026nbsp;1(d)\u003c/b\u003e. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003e shows the entire filtering and calculation process from first-level candidate stops to third-level candidate stops.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo eliminate adjacent stops that are too close in distance, the level three candidate stops are divided and labeled according to the routes. The path distance between adjacent candidate stops on each route is calculated. Since Hohhot's city center has a large passenger flow, a spacing of 100 meters between stops is considered an acceptable minimum distance. Therefore, adjacent candidate stops with less than a 100m interval are compared, and the stop with a higher weight is retained while the stop with a lower weight is eliminated. After filtering, the stops that meet the criteria are selected by the value of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({W}_{j,2}\\)\u003c/span\u003e\u003c/span\u003e, retaining those with higher \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({W}_{j,2}\\)\u003c/span\u003e\u003c/span\u003e and eliminating others, resulting in the desired number of new bus stops.\u003c/p\u003e \u003cp\u003eThese stops represent the newly obtained bus stops in this study, which is equal to the number of actual bus stops within the scope of the research in Hohhot.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eThe Evaluation of Bus Stops by AHP\u003c/h2\u003e \u003cp\u003eThis section describes the evaluation method for both the original bus stops in Hohhot City and the newly obtained bus stops in this study.\u003c/p\u003e \u003cp\u003eThree indicators are used for evaluating the bus stops in this study: the number of POIs within the service range of a bus stop(POI number), the distance between adjacent bus stops(stops distances), and the average distance between the stop and POIs within the stop's service area (stop-POIs distances). The weights for these three indicators are calculated using the Analytic Hierarchy Process (AHP). This method combines quantitative analysis with qualitative analysis by incorporating the decision maker's expertise to determine the relative importance of each evaluation criterion and assign appropriate weightings (J. Li and Zou \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). The results are presented in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e.\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\u003eThe weight of the three indicators\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\u003eIndicator\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePOI number\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003estops distances\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003estop-POI distances\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWeight (C\u003csub\u003ei\u003c/sub\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.63334572\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.260497956\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.106156324\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eA total of 1769 POI records are used for evaluation. The POI dataset includes the POI's ID, type, and coordinates. The types of POIs include schools, shopping malls, parks, companies, government institutions, etc.\u003c/p\u003e \u003cp\u003eThe distribution of numbers of POIs within a 1km radius of bus stops follows an exponential distribution with R-square 0.9914. Bus stops located in the city center have a higher number of POIs, but a lower average clustering coefficient (Zhang et al. \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), indicating a more evenly distributed POI density in the city center. Therefore, it is necessary to assign different weights to bus stops in order to differentiate the importance of POI quantities in different distance categories.\u003c/p\u003e \u003cp\u003ePreliminary calculations reveal that within the study area, there are no more than 60 POIs within a 200m radius of bus stops in Hohhot, and there are very few bus stops with over 30 POIs in their service range. Based on this, for the indicator of POI number, the following scoring criteria are established: a score of 0.3 is assigned if there are 0\u0026ndash;10 POIs within the range, a score of 0.6 is assigned if there are 11\u0026ndash;20 POIs, and a score of 1 is assigned if there are more than 30 POIs.\u003c/p\u003e \u003cp\u003eSince both criteria consist of 64 bus stops, the full score for the 64 bus stops should be 64 points. To convert this proportionally to 10 points, the scoring standards are presented in Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e.\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\u003eThe assignment of POI number\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"2\"\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 \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePOI number range\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAssign points\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e0\u0026ndash;10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.3\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e11\u0026ndash;20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.6\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e21\u0026ndash;30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.8\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFull marks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e64\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStandard full marks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe spacing between bus stops affects the operating speed and service efficiency of buses. As the focus of this study is the city center of Hohhot, an area with inadequate and congested transportation infrastructure, bus speeds are not particularly fast. Therefore, the appropriate distance between adjacent bus stops is determined to be between 100 and 500m (Tirachini \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). The distances between neighboring bus stops on each route are calculated, and scoring is assigned to the bus stops based on distance categories. The specific scores are subjectively determined by the author's years of experience living in Hohhot. The specific scoring criteria are presented in Table\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e.\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\u003eThe assignment of bus spacing\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"2\"\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 \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ebus spacing range\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAssign points\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e0-100m\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.5\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e101-500m\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e501-1000m\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.5\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;1000m\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFull marks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e64\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStandard full marks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eFor the stop-POIs distances indicator, the average distance between each bus stop and all POIs within a 200m radius of the stop is calculated. When this average distance is not more than 100 meters, it indicates that the bus stop is located near several major POIs, which is the ideal situation, earning a score of 1. In all other cases, the attractiveness of the bus stop to passengers slightly decreases. The scores are assigned based on the author's experience living in Hohhot, as shown in Table\u0026nbsp;\u003cspan refid=\"Tab6\" class=\"InternalRef\"\u003e6\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab6\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 6\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eThe assignment of stops-POI distances\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"2\"\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 \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003estop-POI distances range\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAssign points\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e100\u0026ndash;150\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.7\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;150\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.5\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFull marks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e64\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStandard full marks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe scores of the original and new bus stops are separately calculated and normalized in three aspects, ensuring that the scores are distributed within the range of [0, 10]. Then, based on the weights obtained from the AHP (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e), the overall evaluation score is calculated using \u003cb\u003eFormula 3\u003c/b\u003e.\u003cdiv id=\"Equc\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equc\" name=\"EquationSource\"\u003e\n$$\\begin{array}{c}P= \\sum _{l=1}^{3}{C}_{l}\\times {S}_{l}\\#\\left(3\\right)\\end{array}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eP represents the total score, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({C}_{l}\\)\u003c/span\u003e\u003c/span\u003e represents the weights of the three indicators, which respectively indicate the importance of the three indicators, and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({S}_{l}\\)\u003c/span\u003e\u003c/span\u003e represents the scores of the old or new bus stops within the three indicators.\u003c/p\u003e \u003c/div\u003e"},{"header":"RESULTS","content":"\u003cp\u003eThe real 64 bus stops in Hohhot(original stops) are compared with the 64 new bus stops based on the three indicators mentioned above(Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIn terms of POI numbers, the new stops have an advantage over the original stops. The total number of POIs within the service range of the original bus stops is 935, while the total number of POIs within the service range of the new bus stops is 965. Therefore, the new bus stops serve 30 more POIs than the original bus stops. On average, each original bus stop serves 14.609 POIs, while the average number of POIs served by each new bus stop is 15.078. This indicates that each new bus stop serves an additional 0.5 POIs compared to the original bus stops. After scoring, using 10 as the maximum score, the original bus stops achieve a score of 5.406, while the new bus stops achieve a score of 5.594 (Table\u0026nbsp;\u003cspan refid=\"Tab7\" class=\"InternalRef\"\u003e7\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab7\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 7\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eComparison of old and new stops\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\" colspan=\"4\" nameend=\"c4\" namest=\"c1\"\u003e \u003cp\u003ePOI number\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\u003esum\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eaverage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003escores(\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({S}_{1}\\)\u003c/span\u003e\u003c/span\u003e,\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({S}_{1}^{{\\prime }}\\)\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eoriginal stops\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e935\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e14.609\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5.406\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003enew stops\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e965\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e15.078\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5.594\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c4\" namest=\"c1\"\u003e \u003cp\u003ebus stops spacing\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 \u003cp\u003emedian(m)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eaverage(m)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003escores(\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({S}_{2}\\)\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({S}_{2}^{{\\prime }}\\)\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eoriginal stops\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e376.7334646\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e426.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e6.359\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003enew stops\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e308.8874551\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e408.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e6.797\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c4\" namest=\"c1\"\u003e \u003cp\u003ePOI distance\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 \u003cp\u003emedian(m)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eaverage(m)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003escores(\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({S}_{3}\\)\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({S}_{3}^{{\\prime }}\\)\u003c/span\u003e\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eoriginal stops\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e122.245\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e122.536\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e7.328\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003enew stops\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e131.804\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e130.194\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e6.891\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"4\" nameend=\"c4\" namest=\"c1\"\u003e \u003cp\u003eP(final score)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eoriginal stops\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5.8583\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\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\u003enew stops\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e6.0454\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\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 \u003cp\u003eIn terms of bus stop spacing, the average distance between adjacent original bus stops is 426.20 meters, while the average distance between adjacent new bus stops is 408.10 meters. This indicates that the spacing between new bus stops is smaller. The median and mean values of bus stop spacing for both original and new bus stops fall within the normal range, but the spacing between new bus stops generally aligns better with the typical bus stop spacing in urban city centers. The original bus stops achieve a score of 6.359, while the new bus stops achieve a score of 6.797 (Table\u0026nbsp;\u003cspan refid=\"Tab7\" class=\"InternalRef\"\u003e7\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn terms of POI distance, the new stops have a disadvantage compared to the original stops. The average distance between all POIs within the service range of the original stop and the stop itself is 122.536 meters, while the average distance between all POIs within the service range of the new stop and the stop itself is 130.194 meters, which means that the new stop has an average distance of 7.5 meters longer than the original stop; The median score of the original stop is 122.245 meters, and the median score of the new stop is 131.804 meters, indicating that the new stop does not have an advantage in terms of POI distance. After scoring, with a full score of 10, the original stop scored 7.328, and the new stop scored 6.891 (Table\u0026nbsp;\u003cspan refid=\"Tab7\" class=\"InternalRef\"\u003e7\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eBased on the scores of the new and original stops in the above three indicators, the total score can be calculated using the \u003cb\u003eformula 3\u003c/b\u003e. The final total score of bus stops is shown in Table\u0026nbsp;\u003cspan refid=\"Tab7\" class=\"InternalRef\"\u003e7\u003c/span\u003e. The comprehensive score of the original stops in the three indicators is 5.853, and for the new stops, the value is 6.0454. The new stops scored higher and was optimized by 3.194%.\u003c/p\u003e \u003cp\u003eThe conclusions drawn support the rationality of this research method, which suggests that the new stop is better than the original stop in terms of service quality, convenience, and overall user experience. The limited optimization scope may be due to the sparse distribution of POI data within the study area. If the POI data were more densely distributed, the evaluation of the two schemes would likely be more accurate, and the scoring differences between the original and new bus stops would be more pronounced.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eThis paper focuses on 14 bus routes and 64 bus stops within a 4-kilometer radius in the downtown area of Hohhot. Multisource data is used, including POI, the spacing distribution of bus stop and mobile signaling data from residents of Hohhot. The study employs the K-means clustering and the WPDM to calculate new bus stop locations. The new bus stops outperform the existing ones in terms of the POIs number and bus stop spacing, resulting in an overall optimization of 3.194% compared to the original bus stops in Hohhot.\u003c/p\u003e \u003cp\u003eDue to limitations in the data used in this study, the original location data cannot accurately represent the distribution characteristics of bus passenger flows. Consequently, the optimization level of the bus stop selection method in this study is relatively low. Therefore, finding more optimal data sources will be a key focus for future research. Additionally, expanding the geographic scope of the study would enhance its value, and there is room for improvement in the scoring method for evaluation indicators.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eEthical Approval \u0026nbsp;\u003c/h2\u003e\n\u003cp\u003eThis declaration is \u0026ldquo;not applicable\u0026rdquo;.\u003c/p\u003e\n\u003ch2\u003eCompeting interests \u0026nbsp;\u003c/h2\u003e\n\u003cp\u003eNo conflict of interest exits in the submission of this manuscript, and manuscript is approved by all authors for publication.\u003c/p\u003e\n\u003ch2\u003eAuthors\u0026apos; contributions \u0026nbsp;\u003c/h2\u003e\n\u003cp\u003eJY wrote the main manuscript text and prepared all figures. ZY revised the initial draft and modified Table 7. CW and GS collected and organized references. All authors read and approved the final manuscript.\u003c/p\u003e\n\u003ch2\u003eFunding \u0026nbsp;\u003c/h2\u003e\n\u003cp\u003eOur study is supported by Young Scientists Fund of the National Natural Science Foundation of China (Grant No. 61903205) and Young Scientists Fund of Natural Science of Inner Mongolia (Grant No. 2019BS07002).\u003c/p\u003e\n\u003ch2\u003eData Availability Statement \u0026nbsp;\u003c/h2\u003e\n\u003cp\u003eThe cell phone mobility data was acquired from China Unicom, and they have not given their permission for researchers to share their data. The POI data was acquired from Gaode Map, data requests can be made to Gaode Map via this website: https://lbs.amap.com/.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAhmad, F., Iqbal, A., Ashraf, I., \u0026amp; Marzband, M. (2022). Optimal location of electric vehicle charging station and its impact on distribution network: A review. \u003cem\u003eEnergy Reports, 8\u003c/em\u003e, 2314-2333.\u003c/li\u003e\n\u003cli\u003eDanese, A., Tors\u0026aelig;ter, B. N., Sumper, A., \u0026amp; Garau, M. (2022). Planning of high-power charging stations for electric vehicles: A review. \u003cem\u003eApplied Sciences, 12\u003c/em\u003e(7), 3214.\u003c/li\u003e\n\u003cli\u003eGuangyuan, Y. (2018). Evaluation method of current capacity of urban bus stops based on POI data and ArcGIS Spatial analysis technology. \u003cem\u003eBeauty \u0026amp; Times\u003c/em\u003e(4), 67-68.\u003c/li\u003e\n\u003cli\u003eJi-zhuo, L., Ting, X., \u0026amp; Min, Z. (2022). Visualization Analysis of Bus Stop Optimization Based on Multi-source Data. \u003cem\u003eJournal of Chinese Computer Systems, 43\u003c/em\u003e(7).\u003c/li\u003e\n\u003cli\u003eKchaou-Boujelben, M. (2021). Charging station location problem: A comprehensive review on models and solution approaches. \u003cem\u003eTransportation Research Part C: Emerging Technologies\u0026nbsp;\u003c/em\u003e\u003cem\u003e132\u003c/em\u003e, 103376.\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"6\"\u003e\n\u003cli\u003eLi, J., \u0026amp; Zou, P. X. W. (2011). Fuzzy AHP-Based Risk Assessment Methodology for PPP Projects. \u003cem\u003eJournal of Construction Engineering and Management-Asce, 137\u003c/em\u003e(12), 1205-1209. doi:10.1061/(asce)co.1943-7862.0000362\u003c/li\u003e\n\u003cli\u003eLi, P., Jiang, L., Zhang, S., \u0026amp; Jiang, X. (2022). Demand Response Transit Scheduling Research Based on Urban and Rural Transportation Station Optimization. \u003cem\u003eSustainability, 14\u003c/em\u003e(20), 13328.\u003c/li\u003e\n\u003cli\u003eLing, S., Ma, S., \u0026amp; Jia, N. (2022). Sustainable urban transportation development in China: A behavioral perspective. \u003cem\u003eFrontiers of Engineering Management, 9\u003c/em\u003e(1), 16-30.\u003c/li\u003e\n\u003cli\u003ePerumal, S. S., Lusby, R. M., \u0026amp; Larsen, J. (2022). Electric bus planning \u0026amp; scheduling: A review of related problems and methodologies. \u003cem\u003eEuropean Journal of Operational Research, 301\u003c/em\u003e(2), 395-413.\u003c/li\u003e\n\u003cli\u003eShen, Z.-J. M., Feng, B., Mao, C., \u0026amp; Ran, L. (2019). Optimization models for electric vehicle service operations: A literature review. \u003cem\u003eTransportation Research Part B: Methodological\u0026nbsp;\u003c/em\u003e\u003cem\u003e128\u003c/em\u003e, 462-477.\u003c/li\u003e\n\u003c/ol\u003e\n\u003col start=\"11\"\u003e\n\u003cli\u003eSiji, B. (2018). \u003cem\u003eResearch on bus entry queuing probability and station location based on intelligent bus data.\u003c/em\u003e (Master Master), Southwest Jiaotong University, Available from Cnki\u003c/li\u003e\n\u003cli\u003eSyakur, M. A., Khotimah, B. K., Rochman, E. M. S., \u0026amp; Satoto, B. D. (2017, 2018\u003c/li\u003e\n\u003cli\u003eNov09). \u003cem\u003eIntegration K-Means Clustering Method and Elbow Method For Identification of The Best Customer Profile Cluster.\u003c/em\u003e Paper presented at the 2nd International Conference on Vocational Education and Electrical Engineering (ICVEE), Univ Negeri Surabaya, Fac Engn, Dept Elect Engn, Surabaya, INDONESIA.\u003c/li\u003e\n\u003cli\u003eTirachini, A. (2014). The economics and engineering of bus stops: Spacing, design and congestion. \u003cem\u003eTransportation research part A: policy practice 59\u003c/em\u003e, 37-57.\u003c/li\u003e\n\u003cli\u003eWei, C., \u0026amp; Tao, L. (2020). Optimization of bus stops based on multi-source Big data. \u003cem\u003e2020 China Urban Transportation planning Annual Conference\u003c/em\u003e, 1070-1080. doi:10.26914/c.cnkihy.2020.032165\u003c/li\u003e\n\u003cli\u003eXiaoxiong, W., Yongxin, L., \u0026amp; Jukang, L. (2019). Research on big data mining based evaluation method for life service\u003c/li\u003e\n\u003cli\u003elevel of urban public transport stops. \u003cem\u003eModern Electronics Technique, 42\u003c/em\u003e(02), 67-70. doi:10.16652/j.issn.1004⁃373x.2019.02.016\u003c/li\u003e\n\u003cli\u003eXue, F., \u0026amp; Yao, E. (2022). Impact analysis of residential relocation on ownership, usage, and carbon-dioxide emissions of private cars. \u003cem\u003eEnergy Reports, 252\u003c/em\u003e, 124110.\u003c/li\u003e\n\u003cli\u003eYiyi, C., Jianhua, G., \u0026amp; Huanxin, J. (2022). An Evaluation of Bus Station Layout Rationality Using Matter Element\u003c/li\u003e\n\u003cli\u003eAnalysis Based on Points of Interest Data. \u003cem\u003eJournal of Transport Information and Safety, 38\u003c/em\u003e(06), 63-72. doi:10. 3963/j. jssn. 1674-4861. 2020. 06. 009\u003c/li\u003e\n\u003cli\u003eZhang, H., Li, X., Zhang, L., Wang, W., Jia, J., \u0026amp; Shi, B. (2021). Discovering station patterns of urban transit network with multisource data: empirical evidence in Jinan, China. \u003cem\u003eKSCE Journal of Civil Engineering 25\u003c/em\u003e(2), 680-691.\u003c/li\u003e\n\u003cli\u003eZu-peng, L., Ke-ping, L., Ya-qin, H., \u0026amp; Cai-xia, X. (2019). Optimization of Bus Stop Layout Based on Time Distance Trajectory. \u003cem\u003eJournal of Highway and Transportation \u003c/em\u003e\u003cem\u003eＲesearch and Development, 36\u003c/em\u003e(6), 103-111.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"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":"Bus Stops, POI, Clustering, K-means, Spatial Analysis","lastPublishedDoi":"10.21203/rs.3.rs-3913979/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3913979/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe location of bus stops is crucial for attracting bus passengers. Researchers have explored various methods, such as questionnaire surveys and field investigations, to determine the location of bus stops, but these methods largely rely on experience and intuition. With the introduction of big data, a more scientific approach has emerged for siting bus stops. Based on the location data from cellular base stations in Hohhot, China, this paper uses K-means clustering and spatial analysis to calculate the coordinates of bus stops within 4 kilometers of the central business district (CBD) area. Then the study utilizes the city point of interest (POI) data to compare the new bus stops and existing ones in terms of the POI number, bus stop spacing, and POI average distance. The study concludes that the new bus stops outperform the existing ones in terms of POI number and stops spacing. The Analytic Hierarchy Process (AHP) is applied to evaluate the results that indicate the newly constructed bus stops are superior to the existing ones, with an overall optimization rate of 3.194%.The findings of this study can serve as decision-making references for urban planning departments and public transportation operators, aiming to increase public transport passenger flow and improve the traffic conditions in cities. Additionally, the methods utilized in this research can be applied to other cities, assisting them in site selection of bus stops and optimal planning, thereby promoting the development of public transportation.\u003c/p\u003e","manuscriptTitle":"Optimization of Bus Stop Location based on Spatial Analysis and K-Means Clustering Method – A Case Study in Hohhot City","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-02-27 21:01:02","doi":"10.21203/rs.3.rs-3913979/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":"c3d7f4f2-2abf-4dd6-b73a-e1d01fab920b","owner":[],"postedDate":"February 27th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-03-31T18:01:11+00:00","versionOfRecord":[],"versionCreatedAt":"2024-02-27 21:01:02","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-3913979","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3913979","identity":"rs-3913979","version":["v1"]},"buildId":"omnImTCwR2MFx8CMYfrG7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}