Method and device for identifying spatial difference of urban public transport service based on travel behavior data

Abstract

The application discloses a kind of city public transport service space difference identification method and device based on travel behavior data, the method includes: obtaining and preprocessing public transport data and facility data;Comprehensive determination of travel purpose;Group-purpose-period-distance four-dimensional travel image label system is constructed;Multi-dimensional expansion analysis is carried out using singular value decomposition method, and image mode characteristics are extracted;Aggregation calculates image intensity and supply-demand ratio index;Distance centrality index and net flow intensity index are calculated, combined with image intensity and supply-demand ratio index, to construct space service difference identification index system;Each index is standardized, and K-means clustering algorithm is used to identify and summarize various space regions.The application realizes the fine description of travel behavior and the accurate identification of spatial service difference, reveals the structural service mismatch problem, provides a scientific basis for city public transport service optimization, and can be widely applied in urban and rural planning and traffic engineering field.

Description

Technical Field

[0001] This invention relates to a method and apparatus for identifying spatial differences in urban public transport services based on travel behavior data, belonging to the fields of urban and rural planning and transportation engineering. Background Technology

[0002] Public transportation, as a vital public service infrastructure in cities, plays a crucial role in connecting social services, economic activities, and residential areas. In recent years, with the acceleration of urbanization and the increase in motorization, urban public transportation systems have undergone unprecedented expansion. Cities worldwide have invested heavily in expanding network coverage and improving service quality, making the development of integrated public transportation systems an important component of sustainable urban planning.

[0003] However, rapid expansion of transportation networks has not always resulted in equitable service provision. Against the backdrop of an aging population, rapid economic development, and increasingly diversified travel choices, citizens' travel purposes, travel times, travel distances, and the characteristics of different travel groups have all undergone significant changes. For public transportation services, including route design and station location selection, the factors that need to be considered are becoming increasingly complex. Traditional public transportation planning methods often overlook the diverse accessibility needs and differentiated travel patterns among different population groups, leading to a systemic mismatch between service supply and actual demand.

[0004] The methods for measuring public transport fairness and studying spatial supply-demand disparities have evolved from empiricism and qualitative research to spatial interaction models, GIS spatial analysis methods, and big data-based analysis methods. Existing research primarily relies on factors such as population, transportation, facility distribution, and travel costs, but it has the following limitations:

[0005] Existing assessment methods typically rely on general indicators and fail to identify specific service-demand mismatches at the behavioral pattern level. Transportation equity has garnered significant attention because service inequality exacerbates social inequality and limits opportunities for disadvantaged groups. Smart card technology has transformed transportation analytics, providing data on actual travel patterns rather than just travel preferences, and recent methodological developments have expanded to user profiling and behavioral pattern recognition. However, a gap remains in the relationship between behavioral insights and spatial equity assessment; existing research rarely systematically integrates these two dimensions, with most studies focusing either on spatial accessibility analysis or behavioral pattern recognition.

[0006] Traditional assessments often obscure important differences within travel groups. For example, while the elderly population as a whole may exhibit lower travel frequency, there are significant variations in their travel purposes, time preferences, and distance preferences. Simply analyzing "the elderly" as a homogeneous group fails to reveal the precise matching between their segmented needs and service supply.

[0007] Spatial identification based on various dimensions often relies on single-level fairness considerations and lacks a comprehensive evaluation system. Existing studies mostly assess these differences from a single perspective, such as facility accessibility, travel time cost, and service coverage, failing to fully consider the multidimensional heterogeneity of travel behavior and its interaction with spatial service distribution. Currently, there is a need for a comprehensive method that integrates behavioral heterogeneity and spatial difference analysis to accurately identify spatial differences in urban public transport services, providing a scientific basis for refined and targeted public transport service optimization. Summary of the Invention

[0008] In view of this, the present invention provides a method, device, computer equipment and storage medium for identifying spatial differences in urban public transport services based on travel behavior data. It achieves a refined characterization of travel behavior and accurate identification of spatial service differences, reveals the structural service mismatch problem, provides a scientific basis for optimizing urban public transport services, and can be widely applied in the fields of urban and rural planning and transportation engineering.

[0009] The first objective of this invention is to provide a method for identifying spatial differences in urban public transport services based on travel behavior data.

[0010] The second objective of this invention is to provide a device for identifying spatial differences in urban public transport services based on travel behavior data.

[0011] A third objective of this invention is to provide a computer device.

[0012] A fourth objective of this invention is to provide a storage medium.

[0013] The first objective of this invention can be achieved by adopting the following technical solution:

[0014] A method for identifying spatial differences in urban public transport services based on travel behavior data, the method comprising:

[0015] The system acquires public transport data and facility data for the target area, preprocesses the data, and divides the target area into hexagonal grids as basic zoning units. The public transport data includes bus card swipe data, bus stop vector data, and bus route vector data. The bus card swipe data includes user identifier, travel group identifier, travel time, origin station, destination station, and travel route. The bus stop vector data includes station name, station coordinates, and the route to which it belongs. The bus route vector data includes route name and route stop sequence. The facility data includes land use vector data and POI data.

[0016] Based on data on the origin and destination stations and their surrounding facilities, and combined with travel time and distance, the purpose of travel is comprehensively determined.

[0017] Based on travel groups, travel purposes, travel time periods, and travel distances, a four-dimensional travel profile tagging system is constructed to generate a user-profile frequency matrix;

[0018] The user-profile frequency matrix is ​​normalized, and singular value decomposition is used for multidimensional extended analysis to extract profile pattern features.

[0019] Travel records are mapped to a hexagonal grid, and the profile intensity index of each grid is aggregated and calculated. Based on the travel purpose attributes in the profile, demand-oriented profiles and supply-oriented profiles are distinguished, and the supply-demand ratio index of each grid is calculated.

[0020] Calculate the distance centrality index and net flow intensity index, and combine them with the profile intensity index and supply-demand ratio index to construct a spatial service difference identification index system;

[0021] The indicators in the spatial service difference identification index system are standardized, and the standardized indicator data are analyzed using the K-means clustering algorithm to identify and summarize various spatial regions, and output the spatial difference identification results of urban public transport services.

[0022] Furthermore, the preprocessing of public transport data and facility data includes:

[0023] Match the station names in the bus card swipe data with the bus station vector data and remove invalid records;

[0024] The latitude and longitude coordinates of each station are obtained through the geocoding API to supplement the spatial coordinate information of the station vector data;

[0025] By iterating through the origin and destination station pairs of bus card swiping data using the route planning API, a station adjacency matrix is ​​constructed to store the distance information and estimated travel time of all station pairs.

[0026] Filter and reclassify POI data.

[0027] Furthermore, the determination of travel purpose based on data of the origin station, destination station, and surrounding facilities, combined with travel time and distance, includes:

[0028] Using a spatial matching method, the travel destination station is spatially connected with the surrounding land use plots to preliminarily determine the travel function type;

[0029] The POI proximity analysis method is used to calculate the distance attenuation weight from the destination station to each type of POI to help determine the travel purpose; the formula for calculating the distance attenuation weight is as follows:

[0030] ;

[0031] In the formula, i represents the POI type, and j represents the POI of type i at the j-th level. The total number of POIs of type i For the weight of type i, Let j be the Euclidean distance from the destination station to the j-th point of interest (POI). To influence the radius;

[0032] The purpose of travel is determined by combining the characteristics of travel time and travel distance.

[0033] Furthermore, the normalization of the user-profile frequency matrix and the use of singular value decomposition for multidimensional extended analysis to extract profile pattern features include:

[0034] The user-profile frequency matrix is ​​normalized by dividing each row by the user's total number of trips to obtain the profile frequency distribution matrix.

[0035] Singular value decomposition (SVD) was performed on the image frequency distribution matrix to perform SVD separately for the elderly group and the general population. The decomposition formula is as follows:

[0036] ;

[0037] In the formula, Let U be the frequency distribution matrix of the user profile, and U be the user feature matrix, where each row represents a user's score on k potential patterns. It is a singular value diagonal matrix, where the diagonal elements represent the importance of each pattern; is the profile pattern matrix, where each row represents a profile combination corresponding to a potential pattern; k is the number of dimensions after dimensionality reduction.

[0038] After performing singular value decomposition on the elderly group and the general population respectively, the top k components are retained to ensure that the cumulative variance explained is greater than the preset percentage.

[0039] Pattern features are extracted from the retained k components, and the main behavioral patterns are identified and named according to the weight distribution of each component in the profile dimension.

[0040] Furthermore, the process of mapping travel records to a hexagonal grid and aggregating and calculating the profile intensity index of each grid includes:

[0041] Spatial matching of each station's coordinates with a hexagonal grid establishes a station-grid mapping relationship;

[0042] For each travel record, based on the grid where the origin and destination stations are located, the inflow and outflow of that grid are statistically analyzed.

[0043] The intensity of the profiles in each grid is aggregated and statistically analyzed to calculate the intensity of the elderly population, the intensity of the general population, the total activity intensity, the profile diversity index, the population balance, the proportion of the elderly, and the functional specialization index.

[0044] Furthermore, based on the travel purpose attributes in the profile, the process distinguishes between demand-oriented and supply-oriented profiles, and calculates the supply-demand ratio for each grid, including:

[0045] Based on the purpose of travel, the profiles are divided into two categories: demand-oriented profiles and supply-oriented profiles. The demand-oriented profiles include travel purposes such as medical treatment, shopping, education, and leisure, while the supply-oriented profiles include travel purposes such as going home, commuting, and transferring.

[0046] For each grid, the number of demand-oriented trips and the number of supply-oriented trips are counted separately.

[0047] The supply-demand ratio is calculated using the following formula:

[0048] ;

[0049] In the formula, The supply-oriented travel intensity for grid G ​​is defined as the number of trips flowing into grid G. Demand-driven travel intensity for grid G, i.e., the number of trips flowing out of grid G.

[0050] Furthermore, the indicators in the spatial service difference identification indicator system are standardized, and the standardized indicator data are analyzed using the K-means clustering algorithm to identify and summarize various spatial regions, outputting the spatial difference identification results for urban public transport services, including:

[0051] Z-score standardization was applied to the distance centrality index, the proportion of elderly people, the profile diversity index, the population balance index, the supply-demand ratio index, the net mobility intensity index, and the functional specialization index.

[0052] K-means clustering algorithm was used to perform cluster analysis on the standardized index data;

[0053] Based on the clustering results, various spatial regions are identified, including core service areas, balanced development areas, peripheral residential areas, and transitional mixed areas.

[0054] Analyze the distribution of characteristic indicators of various spatial regions, and summarize the basic characteristics and service difference patterns of various spatial regions.

[0055] The second objective of this invention can be achieved by adopting the following technical solution:

[0056] A device for identifying spatial differences in urban public transport services based on travel behavior data, the device comprising:

[0057] The acquisition and preprocessing module is used to acquire public transportation data and facility data of the target area, preprocess the public transportation data and facility data, and divide the target area into hexagonal grids as basic zoning units. The public transportation data includes public transportation card swipe data, public transportation station vector data, and public transportation route vector data. The public transportation card swipe data includes user identifier, travel group identifier, travel time, origin station, destination station, and travel route. The public transportation station vector data includes station name, station coordinates, and the route to which it belongs. The public transportation route vector data includes route name and route station sequence. The facility data includes land use vector data and POI data.

[0058] The determination module is used to comprehensively determine the purpose of travel based on data of the origin station, destination station and surrounding facilities, combined with travel time and distance.

[0059] The first construction module is used to build a four-dimensional travel profile tag system based on travel groups, travel purposes, travel time periods and travel distances, and generate a user-profile frequency matrix;

[0060] The extraction module is used to normalize the user-profile frequency matrix, perform multidimensional extended analysis using singular value decomposition, and extract profile pattern features.

[0061] The calculation module is used to map travel records to a hexagonal grid, aggregate and calculate the profile intensity index of each grid, and distinguish between demand-oriented profiles and supply-oriented profiles based on the travel purpose attributes in the profiles, and calculate the supply-demand ratio index of each grid.

[0062] The second construction module is used to calculate the distance centrality index and the net flow intensity index, and combine them with the profile intensity index and the supply-demand ratio index to construct a spatial service difference identification index system.

[0063] The identification module is used to standardize the various indicators in the spatial service difference identification indicator system, and uses the K-means clustering algorithm to analyze the standardized indicator data, identify and summarize various spatial regions, and output the spatial difference identification results of urban public transport services.

[0064] The third objective of this invention can be achieved by adopting the following technical solution:

[0065] A computer device includes a processor and a memory for storing a processor-executable program, wherein when the processor executes the program stored in the memory, it implements the above-described method for identifying spatial differences in urban public transport services.

[0066] The fourth objective of this invention can be achieved by adopting the following technical solution:

[0067] A storage medium storing a program that, when executed by a processor, implements the above-described method for identifying spatial differences in urban public transport services.

[0068] The present invention has the following advantages over the prior art:

[0069] 1. This invention constructs a four-dimensional travel profile system (group-destination-time period-distance), enabling a refined depiction of travel behavior. Compared to traditional methods that only consider demographic characteristics or single travel attributes, this invention can capture the multidimensional heterogeneity of user travel behavior, thereby more accurately identifying the service needs of different segmented groups.

[0070] 2. This invention uses the singular value decomposition method to reduce the dimensionality and extract patterns from the user-profile frequency matrix, which can identify potential dominant behavior patterns from high-dimensional and complex travel data. By employing a component retention strategy that achieves a cumulative variance explanation rate of more than a preset percentage, the effectiveness and representativeness of pattern extraction are ensured.

[0071] 3. This invention constructs a spatial service difference identification index system encompassing multiple indicators, comprehensively characterizing spatial service differences from multiple dimensions such as distance centrality, group structure, profile diversity, group balance, supply and demand status, flow characteristics, and functional specialization. The K-means clustering algorithm is used to identify four spatial region types, providing a scientific basis for differentiated public transport service optimization strategies. Attached Figure Description

[0072] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0073] Figure 1 This is a simplified flowchart of the urban public transport service spatial difference identification method based on travel behavior data according to Embodiment 1 of the present invention.

[0074] Figure 2 This is a detailed flowchart of the urban public transport service spatial difference identification method based on travel behavior data according to Embodiment 1 of the present invention.

[0075] Figure 3 This is a schematic diagram of the target area and public transportation network in Embodiment 1 of the present invention.

[0076] Figure 4 This is a spatial distribution map of image diversity in Embodiment 1 of the present invention.

[0077] Figure 5This is a spatial distribution diagram of the population balance degree in Embodiment 1 of the present invention.

[0078] Figure 6 This is a spatial distribution map of supply and demand classification in Embodiment 1 of the present invention.

[0079] Figure 7 These are radar images of four types of spatial region features from Embodiment 1 of the present invention.

[0080] Figure 8 This is a spatial distribution map of four types of regions in Embodiment 1 of the present invention.

[0081] Figure 9 This is a structural block diagram of the urban public transport service spatial difference identification device based on travel behavior data according to Embodiment 2 of the present invention.

[0082] Figure 10 This is a structural block diagram of the computer device according to Embodiment 3 of the present invention. Detailed Implementation

[0083] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0084] Example 1:

[0085] like Figure 1 and Figure 2 As shown in the figure, this embodiment provides a method for identifying spatial differences in urban public transport services based on travel behavior data. The method includes the following steps:

[0086] S201. Obtain public transport data and facility data for the target area, preprocess the public transport data and facility data, and divide the target area into hexagonal grids as basic zoning units.

[0087] This embodiment focuses on Beijing as the target area, covering the core urban area within the Sixth Ring Road, an area of ​​approximately 2,000 square kilometers, with a resident population of over 21 million, of which the elderly population accounts for 23.5%. Beijing's public transportation system is one of the world's largest public transportation networks, with over 16,000 bus stops and 1,200 routes.

[0088] The public transport data in this embodiment includes public transport card swiping data, public transport stop vector data, and public transport route vector data. Specifically, it acquires Beijing public transport smart card swiping data from December 17th to 18th, 2021 (working days), containing approximately 3.5 million travel records. The data fields include: physical card number (unique user identifier), card type (distinguishing between senior citizen cards and regular cards), swiping time (travel time), boarding stop (travel origin), alighting stop (travel destination), and travel route number. Among these, there are approximately 520,000 senior citizen card users and approximately 1.85 million regular card users. Vector data of Beijing public transport stops and routes is acquired from the OpenStreetMap (OSM) open platform. The stop data includes attribute information such as stop name and the route it belongs to; the route data includes information such as route name and stop sequence.

[0089] Preprocessing of bus card swiping data: Delete records with missing key fields; delete records with the same boarding and alighting stops; delete records with abnormal card swiping times; label user groups (elderly / general public) according to card type.

[0090] Preprocessing of bus stop vector data: Matching stop names in card swipe data with vector data; for cases where names are inconsistent but the stops are actually the same, standardizing the names and removing cancelled or renamed stop data, and renumbering the stops uniformly; using the Geocoding API service of the Gaode Map Open Platform, a Python program is written to call the API interface to obtain the latitude and longitude coordinates of each stop. The API call parameters are set as follows: city: "Beijing", output: "JSON", and the returned field: location (latitude and longitude); the obtained coordinate information is added to the stop vector data to form complete stop spatial data.

[0091] Constructing a station adjacency matrix: Using the route planning service of the Gaode Map Open Platform, traverse all origin and destination station pairs in the bus card swipe data to obtain the distance and travel time information between stations. API call parameters are set as follows: origin: origin station coordinates; destination: destination station coordinates; strategy: 0 (fastest mode); city: "Beijing"; returned fields: distance (distance in meters), duration (time in seconds); construct a station adjacency matrix with N×N dimensions, where N is the total number of stations. Matrix element (i,j) stores the distance and time information from station i to station j. This distance information is added to the bus card swipe data as the data basis for the travel distance dimension.

[0092] The facility data in this embodiment includes land use vector data and POI data. Specifically, the land use vector data is obtained as follows: land use vector data of Beijing is obtained from the publicly available dataset "Mapping of Basic Urban Land Use Types in China: Preliminary Results in 2018" (EULUC-China); using ArcGIS software, land use vector data within the Sixth Ring Road is extracted according to the boundary of the target area using the intersection tool, and necessary projection transformation and topology checks are performed; POI (Point of Interest) data within the study area is obtained from the Gaode Map Open Platform. The POI data includes information such as coordinates, name, and type.

[0093] Preprocessing of POI data: Remove POI types that are irrelevant to the travel purpose (such as place name and address types); reclassify POIs into 7 categories: catering, medical and health, life services, shopping, leisure and entertainment, residential and business, as shown in Table 1 below; retain the coordinates, specific type and name information of POIs.

[0094] Table 1. POI Functional Reclassification Results

[0095]

[0096] In this embodiment, the target area is divided into hexagonal grids as the basic zoning units, such as... Figure 3 As shown, it has the following advantages: each hexagon is equidistant from its adjacent hexagons, resulting in better isotropy; the hexagon has the fewest sides, reducing edge effects; and the hexagon is closer to a circle, which can better simulate the service coverage area.

[0097] The hexagonal grid generation parameters in this embodiment are set as follows: radius (circumcircle radius): 1000 meters; grid type: regular hexagon; projected coordinate system: WGS84 UTM Zone 50N; the hexagonal grid is generated using Python's H3 geospatial index library, covering the entire study area. A unique grid ID is assigned to each hexagon, and the coordinates of the center point of the hexagon are calculated. Approximately 2000 hexagonal grids are generated and saved as GeoJSON format files.

[0098] S202. Based on data of the origin station, destination station and surrounding facilities, and combined with travel time and distance, the purpose of travel is comprehensively determined.

[0099] Furthermore, step S202 specifically includes:

[0100] S2021. Using a spatial matching method, the travel destination station is spatially connected with the surrounding land use plots to preliminarily determine the travel function type.

[0101] This embodiment uses ArcGIS software to spatially connect travel destinations with land use plots. For each travel record: a 500-meter buffer zone is created around the destination; the area ratio of each type of land use within the buffer zone is calculated; and the travel function is preliminarily determined based on the dominant land use type.

[0102] In this embodiment, the mapping relationship between land use and travel purpose is as follows: residential land maps to the purpose of returning home; commercial land maps to the purpose of shopping; office land maps to the purpose of commuting; educational land maps to the purpose of education; medical land maps to the purpose of seeking medical treatment; and park green space maps to the purpose of leisure.

[0103] S2022. Using the POI proximity analysis method, calculate the distance attenuation weight from the destination station to each type of POI to help determine the travel purpose.

[0104] This embodiment calculates the distance attenuation weight from the trip destination to each type of POI. For the trip destination s of trip record r, the distance attenuation weight from it to each POI is calculated as follows:

[0105] ;

[0106] In the formula, i represents the POI type (dining, medical, shopping, etc.), and j represents the POI of type i at the j-th level. The total number of POIs of type i For the weight of type i, Let be the Euclidean distance from the destination station s to the j-th POI. To minimize the impact radius, it is set to 500 meters.

[0107] Calculate the cumulative weight of each type of POI, and use the type with the highest weight as the result of the POI proximity analysis.

[0108] S2023. Combine travel time characteristics and travel distance characteristics to comprehensively determine the purpose of travel.

[0109] The purpose of travel can be determined by considering travel time and distance characteristics. Travel time characteristics are as follows: Morning peak (6-9 am) combined with long distance indicates a commuting tendency; evening peak (5-8 pm) combined with long distance indicates a return-home tendency; off-peak (10-4 pm) combined with short distance and medical POI destinations indicates a medical-seeking tendency; off-peak (10-4 pm) combined with short distance and commercial POs indicates a shopping tendency; off-peak (10-4 pm) combined with parks and green spaces indicates a leisure tendency.

[0110] The results of the triple verification are comprehensively judged: if the judgment results of the three methods are consistent, the judgment result is adopted; if the judgment results of the three methods are inconsistent, the judgment is made according to the priority of POI proximity, spatial matching, and temporal features; for cases that cannot be clearly judged, they are marked as "other"; finally, the travel purpose is divided into 8 categories: commuting, going home, medical treatment, shopping, leisure, education, transfer, and other; the recognition results are added to the public transport card swiping data.

[0111] Based on actual data from Beijing, the distribution of travel destinations is shown in Table 2 below.

[0112] Table 2 Distribution of Public Transportation Trips in Beijing

[0113]

[0114] S203. Based on the travel group, travel purpose, travel time and travel distance, construct a four-dimensional travel profile tag system and generate a user-profile frequency matrix.

[0115] The four dimensions of travel group, travel purpose, travel time, and travel distance are defined as follows:

[0116] 1) Group dimension: elderly group, general public (2 categories);

[0117] 2) Purpose dimension: commuting, returning home, medical treatment, shopping, leisure, education, transfers, other;

[0118] 3) Time period: Morning peak (6-9 am); Evening peak: 5-8 pm; Off-peak: 10-4 pm; Night: other time periods (4 categories);

[0119] 4) Distance dimension: Short distance: <5 km, medium-short distance: 5-15 km, medium-long distance: 15-30 km, long distance: >30 km (4 categories).

[0120] Theoretically, combining these four dimensions can generate 2 × 8 × 4 × 4 = 256 possible profile combinations. However, in actual travel data, many combinations occur very infrequently. To ensure the representativeness and operability of the analysis, the following steps are used to filter high-frequency profiles:

[0121] 1) Count the frequency of occurrence of each image combination;

[0122] 2) Sort by frequency from highest to lowest;

[0123] 3) Calculate the cumulative coverage rate and select the top N types of images with a coverage rate of 85%;

[0124] 4) Classify the remaining low-frequency profiles into the "Other" category;

[0125] According to statistics from 3.5 million travel records in Beijing, the cumulative coverage of the top 30 high-frequency profiles reached 85.3% (exceeding the preset percentage of 85%). Therefore, this embodiment selects the top 30 profiles as the analysis object.

[0126] The naming rules for profile tags are: group-purpose-time period-distance. For example, "elderly-shopping-off-peak-short trip" indicates a profile of elderly people making short shopping trips during off-peak hours.

[0127] In this embodiment, a user-profile frequency matrix M is generated based on the four-dimensional travel profile tagging system. The matrix has dimensions of m×n, where m is the total number of users and n is the number of profile types (n=30 in this embodiment). The matrix element M[i,j] represents the number of times user i travels with profile type j.

[0128] The process of constructing the user-profile frequency matrix M in this embodiment is as follows: traverse all travel records and generate profile tags based on the four-dimensional attributes; for each user, count the number of trips for each profile type; fill the matrix elements M[i,j]; for profile types that do not appear for the user, fill with 0.

[0129] According to data from Beijing, the user profile matrix for the elderly group has 520,000 dimensions × 30, while the user profile matrix for the general public has 1,850,000 dimensions × 30.

[0130] S204. Normalize the user-profile frequency matrix, use singular value decomposition for multidimensional extended analysis, and extract profile pattern features.

[0131] Furthermore, step S204 specifically includes:

[0132] S2041. Normalize the user-profile frequency matrix by dividing each row by the user's total number of trips, converting the absolute frequency into a frequency distribution (probabilistic), and obtain the profile frequency distribution matrix as follows:

[0133] ;

[0134] The purpose of normalization is to eliminate the impact of differences in the total number of trips among different users, so that the matrix reflects the user profile and preference distribution rather than the absolute amount of activity.

[0135] S2042. Perform singular value decomposition on the image frequency distribution matrix to perform singular value decomposition on the elderly group and the general population respectively.

[0136] To capture specific behavioral patterns within a group, singular value decomposition (SVD) was performed on the normalized matrices of both the elderly and general populations. The decomposition formulas are as follows:

[0137] ;

[0138] In the formula, M is the image frequency distribution matrix, and U is the user feature matrix (m×k), where each row represents a user's score on k potential patterns; It is a singular value diagonal matrix (k×k), where the diagonal elements represent the importance of each pattern; The profile pattern matrix is ​​(k×n), where each row represents a profile combination corresponding to a potential pattern; k is the number of dimensions after dimensionality reduction, and the value of k ranges from 10 to 20; the singular value decomposition parameter is set as follows: the initial value of k is set to 15; the algorithm is TruncatedSVD (scikit-learn library); the random seed is 42.

[0139] S2043. After performing singular value decomposition on the elderly group and the general population respectively, retain the top k components so that the cumulative variance explained is greater than the preset percentage.

[0140] Calculate the variance explained and cumulative variance explained for each component:

[0141] ;

[0142] ;

[0143] Retaining the first k components, the cumulative variance explained ratio is [percentage missing]. (The preset percentage is 85%), and the k value will be dynamically adjusted based on the actual analysis results: if k=15, If k=20, then increase the value of k to 20; if k=10, then decrease the value of k to 10.

[0144] Based on data from Beijing, the actual analysis results are as follows:

[0145] SVD results for the elderly population: number of retained components k=15; cumulative variance explained: 89.7%.

[0146] SVD results for the general population: number of retained components k=15; cumulative variance explained: 89.1%.

[0147] The cumulative variance explained by both groups exceeded the target of 85%, thus meeting the analytical requirements.

[0148] S2044. Extract pattern features from the retained k components, and identify and name the main behavioral patterns based on the weight distribution of each component in the profile dimension.

[0149] In this embodiment, pattern feature analysis is performed on the retained k components. For each component i (corresponding to the i-th row), the top 5 images with the largest weights are extracted, and their common features are analyzed: the group, purpose, time period, and distance attributes of the main images are extracted; the distribution characteristics of each attribute are statistically analyzed; and the pattern is named according to the dominant feature.

[0150] Based on data from Beijing, the seven main behavioral patterns identified are shown in Table 3 below.

[0151] Table 3. Main Behavioral Patterns Identified

[0152]

[0153] The results of singular value decomposition analysis and the feature extraction results of the profile pattern are saved as CSV files for subsequent spatial analysis.

[0154] S205. Map travel records to a hexagonal grid, aggregate and calculate the profile intensity index of each grid, and distinguish between demand-oriented profiles and supply-oriented profiles based on the travel purpose attributes in the profiles, and calculate the supply-demand ratio index of each grid.

[0155] In this embodiment, travel records are mapped to a hexagonal grid, and the image intensity index of each grid is aggregated and calculated, specifically including:

[0156] S2051. Spatial matching of the coordinates of each station with the hexagonal grid to establish the station-grid mapping relationship.

[0157] Specifically, the spatial join method is used: for each station, it is determined whether its coordinates fall within a certain hexagonal grid; if the station falls within the grid, a mapping relationship is established; for stations falling on the grid boundary, they are assigned to the nearest hexagon; spatial matching is implemented using Python's Geopandas library, and spatial indexing is used to accelerate queries. A station-grid mapping table is generated, containing fields such as station ID, station name, and grid ID.

[0158] According to data from Beijing, more than 16,000 bus stops were matched with approximately 1,800 effective hexagonal grids (some edge grids have no stop coverage).

[0159] S2052. For each travel record, based on the grid where the origin and destination stations are located, calculate the inflow and outflow of that grid.

[0160] This embodiment iterates through each travel record and, based on the grids where the origin and destination stations are located, counts the inflow and outflow of each grid: For travel record r, in grid G1 where the origin station s1 is located, the travel generates one "outflow" (demand generation) to G1; in grid G2 where the destination station s2 is located, the travel generates one "inflow" (service provision) to G2; for each grid G, the total number of outflows is counted. Total inflow Net flow Further subdivided statistics by group and profile type: Outflow of elderly population Outflow of the general public Inflow of portrait j The outflow of image j .

[0161] S2053. Perform aggregated statistics on the profile intensity of each grid, and calculate the intensity of the elderly group, the intensity of the general population, the total activity intensity, the profile diversity index, the group balance, and the proportion of the elderly.

[0162] In this embodiment, the following image intensity index is calculated for each hexagonal grid:

[0163] 1) Intensity of group activities:

[0164] Strength of the elderly population: ;

[0165] Mass population strength: ;

[0166] Total activity intensity: .

[0167] 2) Portrait Diversity Index:

[0168] Number of different image types appearing within the grid The spatial distribution of image diversity in this embodiment is as follows: Figure 4 As shown.

[0169] 3) Group balance:

[0170] ;

[0171] The value ranges from [0,1]. The closer the value is to 1, the more balanced the activities of the two groups are. The spatial distribution of the group balance degree in this embodiment is as follows: Figure 5 As shown.

[0172] 4) Proportion of elderly:

[0173] ;

[0174] 5) Functional Specialization Index:

[0175] Calculate based on the purpose type of the dominant image within the grid. A larger value indicates that the grid function is more singular and the degree of specialization is higher.

[0176] For grids with activity intensity below the threshold (total number of trips < 50), exclude them from the analysis scope to avoid the influence of noise.

[0177] In this embodiment, based on the travel purpose attribute in the profile, demand-oriented profiles and supply-oriented profiles are distinguished, and the supply-demand ratio index for each grid is calculated, specifically including:

[0178] S2054. Based on the purpose of travel, the profiles are divided into two categories: demand-oriented profiles and supply-oriented profiles.

[0179] The demand-oriented profile in this embodiment represents the generation of travel demand and the corresponding consumption of service facilities, including: 1) Medical treatment: going to medical institutions to receive services; 2) Shopping: going to commercial areas to consume; 3) Education: going to educational institutions to study; 4) Leisure: going to leisure places for entertainment.

[0180] The supply-oriented profile in this embodiment represents the provision of travel supply, corresponding to the end of residential and employment activities, including: 1) Going home: returning to the place of residence; 2) Commuting: going to the workplace; 3) Transferring: going to a transportation hub.

[0181] It should be noted that "other" travel categories are not classified by supply and demand and are excluded from the calculation.

[0182] S2055. For each grid, count the number of demand-oriented trips and the number of supply-oriented trips respectively.

[0183] For each grid G, calculate the demand intensity:

[0184] ;

[0185] That is, the number of demand-driven trips flowing out of grid G ​​(the generation of demand).

[0186] For each grid G, calculate the supply intensity:

[0187] ;

[0188] That is, the number of supply-oriented trips (service provision) flowing into grid G.

[0189] Further segmentation by group: Demand intensity of the elderly group Supply intensity for the elderly population Demand intensity of the general public Supply intensity of the general public .

[0190] S2056. Calculate the Supply-Demand Ratio (SDR) using the following formula:

[0191] ;

[0192] In the formula, The supply-oriented travel intensity for grid G ​​is defined as the number of trips flowing into grid G. Demand-driven travel intensity for grid G, i.e., the number of trips flowing out of grid G.

[0193] Calculate the supply-demand ratio by group:

[0194] Supply and demand ratio of the elderly population ;

[0195] Supply and demand ratio for the general public: .

[0196] Spatial distribution of supply and demand categories, such as Figure 6 As shown.

[0197] S206. Calculate the distance centrality index and net flow intensity index, and combine them with the profile intensity index and supply-demand ratio index to construct a spatial service difference identification index system.

[0198] 1) Distance to Center (DC): Calculates the shortest path distance from the grid center point to the city center (Beijing Municipal Government), using Dijkstra's shortest path algorithm based on road network data.

[0199] ;

[0200] 2) Elderly Ratio (ER): Calculated in step S205:

[0201] ;

[0202] 3) Portrait Diversity (PD): Calculated in step S205:

[0203] ;

[0204] 4) Group Balance (GB): Calculated in step S205:

[0205] ;

[0206] 5) Supply-Demand Ratio (SDR): Calculated in step S205:

[0207] ;

[0208] 6) Net Flow Intensity (NFI):

[0209] ;

[0210] The value ranges from [0,1]. The larger the value, the more unbalanced the inflow and outflow of the grid.

[0211] 7) Functional Specialization Index (FSI): Calculated in step S205.

[0212] .

[0213] S207. Standardize the various indicators in the spatial service difference identification indicator system, use the K-means clustering algorithm to analyze the standardized indicator data, identify and summarize various spatial regions, and output the spatial difference identification results of urban public transport services.

[0214] Furthermore, step S207 specifically includes:

[0215] S2071. Z-score standardization is performed on the distance centrality index, the proportion of elderly people, the profile diversity index, the population balance index, the supply-demand ratio index, the net mobility intensity index, and the functional specialization index.

[0216] This embodiment uses the Z-score standardization method to standardize the seven indicators:

[0217] ;

[0218] In the formula, This represents the original index value for the i-th grid. This is the mean of the indicator across all grids; This represents the standard deviation of the indicator across all grids; the purpose of standardization is to eliminate the influence of different indicator dimensions and numerical ranges, making each indicator comparable in cluster analysis.

[0219] S2072. The K-means clustering algorithm is used to perform cluster analysis on the standardized index data.

[0220] The K-means clustering algorithm works by assigning each data point (point) to the cluster represented by its nearest cluster centroid, given a value of K and K initial cluster centroids. After all points are assigned, the centroid of that cluster is recalculated based on the average value of all points within the cluster. This process of assigning points and updating cluster centroids is then iteratively repeated until the change in cluster centroids is minimal or a specified number of iterations is reached.

[0221] This embodiment uses the K-means clustering algorithm to perform cluster analysis on the standardized 7-dimensional feature vectors. The algorithm parameters are set as follows: number of clusters K=4; maximum number of iterations is 300; random seed is 42; and K-means clustering is implemented using the scikit-learn library in Python.

[0222] S2073. Based on the clustering results, identify various spatial regions.

[0223] This embodiment matches the clustering results with a grid, analyzes the distribution of characteristic indicators for each cluster, and identifies four types of spatial regions: core service area, balanced development area, peripheral residential area, and transitional mixed area. Based on data from Beijing, the clustering results are shown in Table 4 below, and the radar maps of the four spatial regions are shown below. Figure 7 As shown.

[0224] Table 4 Comparison of Characteristics of Four Types of Spatial Regions

[0225]

[0226] S2074. Analyze the distribution of characteristic indicators of various spatial regions, and summarize the basic characteristics and service difference patterns of various spatial regions.

[0227] The clustering results were imported into ArcGIS software for spatial visualization. Specifically, cluster labels were spatially linked with grid data; different colors were used to represent four region types; and bus stop and route data were overlaid to generate a spatial distribution map, such as... Figure 8 As shown.

[0228] In terms of spatial distribution: Type 1 is concentrated in the central urban area, presenting a core-periphery concentric structure; Type 2 is continuously distributed along the main development axis and secondary centers; Type 3 is mainly distributed on the outskirts of the city, forming obvious residential concentric circles; Type 4 is distributed in a strip or point shape, located in the transition zone between Type 1 / 2 and Type 3.

[0229] It should be noted that although the method operations of the above embodiments are described in a specific order, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. On the contrary, the described steps may be performed in a different order. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0230] Example 2:

[0231] like Figure 9As shown, this embodiment provides a spatial difference identification device for urban public transport services based on travel behavior data. The device includes an acquisition and preprocessing module 901, a judgment module 902, a first construction module 903, an extraction module 904, a calculation module 905, a second construction module 906, and an identification module 907. The specific functions of each module are as follows:

[0232] The acquisition and preprocessing module 901 is used to acquire public transport data and facility data of the target area, preprocess the public transport data and facility data, and divide the target area into hexagonal grids as basic zoning units. The public transport data includes public transport card swipe data, public transport station vector data, and public transport route vector data. The public transport card swipe data includes user identifier, travel group identifier, travel time, origin station, destination station, and travel route. The public transport station vector data includes station name, station coordinates, and the route to which it belongs. The public transport route vector data includes route name and route station sequence. The facility data includes land use vector data and POI data.

[0233] The determination module 902 is used to comprehensively determine the purpose of travel based on the origin station, destination station and surrounding facilities data, combined with travel time and travel distance.

[0234] The first construction module 903 is used to build a four-dimensional travel profile tag system based on travel groups, travel purposes, travel time periods and travel distances, and generate a user-profile frequency matrix;

[0235] Extraction module 904 is used to normalize the user-profile frequency matrix, perform multidimensional extended analysis using singular value decomposition, and extract profile pattern features.

[0236] The calculation module 905 is used to map travel records to a hexagonal grid, aggregate and calculate the profile intensity index of each grid, and distinguish between demand-oriented profiles and supply-oriented profiles based on the travel purpose attributes in the profiles, and calculate the supply-demand ratio index of each grid.

[0237] The second construction module 906 is used to calculate the distance centrality index and the net flow intensity index, and combine them with the profile intensity index and the supply-demand ratio index to construct a spatial service difference identification index system.

[0238] The identification module 907 is used to standardize the various indicators in the spatial service difference identification indicator system, and uses the K-means clustering algorithm to analyze the standardized indicator data, identify and summarize various spatial regions, and output the spatial difference identification results of urban public transport services.

[0239] It should be noted that the device provided in this embodiment is only an example of the above-described division of functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure can be divided into different functional modules to complete all or part of the functions described above.

[0240] Example 3:

[0241] This embodiment provides a computer device, such as... Figure 10 As shown, it includes a processor 1002, a memory, an input device 1003, a display device 1004, and a network interface 1005 connected via a system bus 1001. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium 1006 and internal memory 1007. The non-volatile storage medium 1006 stores an operating system, computer programs, and a database. The internal memory 1007 provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. When the processor 1002 executes the computer programs stored in the memory, it implements the urban public transport service spatial difference identification method of Embodiment 1 above, as follows:

[0242] The system acquires public transportation and facility data for the target area, preprocesses the data, and divides the target area into hexagonal grids as basic zoning units. The public transportation data includes bus card swipe data, bus stop vector data, and bus route vector data. The bus card swipe data includes user identifier, travel group identifier, travel time, origin station, destination station, and route. The bus stop vector data includes station name, station coordinates, and the route it belongs to. The bus route vector data includes route name and route stop sequence. The facility data includes land use vector data and POI data. Based on the origin station, destination station, and surrounding facility data, combined with travel time and distance, the system comprehensively determines the travel purpose. The system then considers the travel group, travel purpose, travel time, and departure time. Based on travel distance, a four-dimensional travel profile labeling system is constructed, generating a user-profile frequency matrix. The user-profile frequency matrix is ​​normalized, and singular value decomposition is used for multi-dimensional extended analysis to extract profile pattern features. Travel records are mapped to a hexagonal grid, and profile intensity indices for each grid are aggregated and calculated. Based on the travel purpose attributes in the profiles, demand-oriented and supply-oriented profiles are distinguished, and the supply-demand ratio indices for each grid are calculated. Distance centrality and net flow intensity indices are calculated. Combining profile intensity and supply-demand ratio indices, a spatial service difference identification index system is constructed. The various indices in the spatial service difference identification index system are standardized, and K-means clustering is used to analyze the standardized index data, identifying and summarizing various spatial regions, and outputting the spatial difference identification results for urban public transport services.

[0243] Example 4:

[0244] This embodiment provides a storage medium, which is a computer-readable storage medium, storing a computer program. When the computer program is executed by a processor, it implements the urban public transport service spatial difference identification method of Embodiment 1 above, as follows:

[0245] The system acquires public transportation and facility data for the target area, preprocesses the data, and divides the target area into hexagonal grids as basic zoning units. The public transportation data includes bus card swipe data, bus stop vector data, and bus route vector data. The bus card swipe data includes user identifier, travel group identifier, travel time, origin station, destination station, and route. The bus stop vector data includes station name, station coordinates, and the route it belongs to. The bus route vector data includes route name and route stop sequence. The facility data includes land use vector data and POI data. Based on the origin station, destination station, and surrounding facility data, combined with travel time and distance, the system comprehensively determines the travel purpose. The system then considers the travel group, travel purpose, travel time, and departure time. Based on travel distance, a four-dimensional travel profile labeling system is constructed, generating a user-profile frequency matrix. The user-profile frequency matrix is ​​normalized, and singular value decomposition is used for multi-dimensional extended analysis to extract profile pattern features. Travel records are mapped to a hexagonal grid, and profile intensity indices for each grid are aggregated and calculated. Based on the travel purpose attributes in the profiles, demand-oriented and supply-oriented profiles are distinguished, and the supply-demand ratio indices for each grid are calculated. Distance centrality and net flow intensity indices are calculated. Combining profile intensity and supply-demand ratio indices, a spatial service difference identification index system is constructed. The various indices in the spatial service difference identification index system are standardized, and K-means clustering is used to analyze the standardized index data, identifying and summarizing various spatial regions, and outputting the spatial difference identification results for urban public transport services.

[0246] It should be noted that the computer-readable storage medium in this embodiment can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof.

[0247] In this embodiment, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this embodiment, the computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable program. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable storage medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable storage medium can be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination thereof.

[0248] The computer-readable storage medium described above can be used to write computer programs for executing this embodiment in one or more programming languages ​​or combinations thereof. These programming languages ​​include object-oriented programming languages—such as Java, Python, and C++—and conventional procedural programming languages—such as C or similar programming languages. The program can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0249] In summary, this invention achieves precise identification from overall assessment to behavioral pattern level, revealing specific service gaps in segmented profiles such as "elderly-medical-off-peak-short-distance"; it constructs a multi-dimensional comprehensive indicator system to comprehensively depict spatial service differences from multiple perspectives such as behavior, groups, functions, and supply and demand; based on real data, it has strong operability and scalability, and can provide a reference for optimizing public transportation services in other cities.

[0250] The above description is merely a preferred embodiment of the present invention. It will be apparent to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects. The scope of the present invention is defined by the appended claims rather than the foregoing description, and thus all variations falling within the meaning and scope of the equivalents of the claims are intended to be included within the present invention.

Claims

1. A method for identifying spatial differences in urban public transport services based on travel behavior data, characterized in that, The method includes: The system acquires public transport data and facility data for the target area, preprocesses the data, and divides the target area into hexagonal grids as basic zoning units. The public transport data includes bus card swipe data, bus stop vector data, and bus route vector data. The bus card swipe data includes user identifier, travel group identifier, travel time, origin station, destination station, and travel route. The bus stop vector data includes station name, station coordinates, and the route to which it belongs. The bus route vector data includes route name and route stop sequence. The facility data includes land use vector data and POI data. Based on data on the origin and destination stations and their surrounding facilities, and combined with travel time and distance, the purpose of travel is comprehensively determined. Based on travel groups, travel purposes, travel time periods, and travel distances, a four-dimensional travel profile tagging system is constructed to generate a user-profile frequency matrix; The user-profile frequency matrix is ​​normalized, and singular value decomposition is used for multidimensional extended analysis to extract profile pattern features. Travel records are mapped to a hexagonal grid, and the profile intensity index of each grid is aggregated and calculated. Based on the travel purpose attributes in the profile, demand-oriented profiles and supply-oriented profiles are distinguished, and the supply-demand ratio index of each grid is calculated. Calculate the distance centrality index and net flow intensity index, and combine them with the profile intensity index and supply-demand ratio index to construct a spatial service difference identification index system; The indicators in the spatial service difference identification index system are standardized, and the standardized indicator data are analyzed using the K-means clustering algorithm to identify and summarize various spatial regions, and output the spatial difference identification results of urban public transport services.

2. The method for identifying spatial differences in urban public transport services according to claim 1, characterized in that, The preprocessing of public transport data and facility data includes: Match the station names in the bus card swipe data with the bus station vector data and remove invalid records; The latitude and longitude coordinates of each station are obtained through the geocoding API to supplement the spatial coordinate information of the station vector data; By iterating through the origin and destination station pairs of bus card swiping data using the route planning API, a station adjacency matrix is ​​constructed to store the distance information and estimated travel time of all station pairs. Filter and reclassify POI data.

3. The method for identifying spatial differences in urban public transport services according to claim 1, characterized in that, The determination of travel purpose based on data of origin and destination stations and their surrounding facilities, combined with travel time and distance, includes: Using a spatial matching method, the travel destination station is spatially connected with the surrounding land use plots to preliminarily determine the travel function type; The POI proximity analysis method is used to calculate the distance attenuation weight from the destination station to each type of POI to help determine the travel purpose; the formula for calculating the distance attenuation weight is as follows: ； In the formula, i represents the POI type, and j represents the POI of type i at the j-th level. The total number of POIs of type i For the weight of type i, Let j be the Euclidean distance from the destination station to the j-th point of interest (POI). To influence the radius; The purpose of travel is determined by combining the characteristics of travel time and travel distance.

4. The method for identifying spatial differences in urban public transport services according to claim 1, characterized in that, The normalization of the user-profile frequency matrix and the use of singular value decomposition for multidimensional extended analysis to extract profile pattern features include: The user-profile frequency matrix is ​​normalized by dividing each row by the user's total number of trips to obtain the profile frequency distribution matrix. Singular value decomposition (SVD) was performed on the image frequency distribution matrix to perform SVD separately for the elderly group and the general population. The decomposition formula is as follows: ； In the formula, Let U be the frequency distribution matrix of the user profile, and U be the user feature matrix, where each row represents a user's score on k potential patterns. It is a singular value diagonal matrix, where the diagonal elements represent the importance of each pattern; is the profile pattern matrix, where each row represents a profile combination corresponding to a potential pattern; k is the number of dimensions after dimensionality reduction. After performing singular value decomposition on the elderly group and the general population respectively, the top k components are retained to ensure that the cumulative variance explained is greater than the preset percentage. Pattern features are extracted from the retained k components, and the main behavioral patterns are identified and named according to the weight distribution of each component in the profile dimension.

5. The method for identifying spatial differences in urban public transport services according to claim 1, characterized in that, The process of mapping travel records to a hexagonal grid and aggregating and calculating the profile intensity index of each grid includes: Spatial matching of each station's coordinates with a hexagonal grid establishes a station-grid mapping relationship; For each travel record, based on the grid where the origin and destination stations are located, the inflow and outflow of that grid are statistically analyzed. The intensity of the profiles in each grid is aggregated and statistically analyzed to calculate the intensity of the elderly population, the intensity of the general population, the total activity intensity, the profile diversity index, the population balance, the proportion of the elderly, and the functional specialization index.

6. The method for identifying spatial differences in urban public transport services according to claim 1, characterized in that, The process of distinguishing between demand-oriented and supply-oriented profiles based on travel purpose attributes in the profile, and calculating the supply-demand ratio for each grid, includes: Based on the purpose of travel, the profiles are divided into two categories: demand-oriented profiles and supply-oriented profiles. The demand-oriented profiles include travel purposes such as medical treatment, shopping, education, and leisure, while the supply-oriented profiles include travel purposes such as going home, commuting, and transferring. For each grid, the number of demand-oriented trips and the number of supply-oriented trips are counted separately. The supply-demand ratio is calculated using the following formula: ； In the formula, The supply-oriented travel intensity for grid G ​​is defined as the number of trips flowing into grid G. Demand-driven travel intensity for grid G, i.e., the number of trips flowing out of grid G.

7. The method for identifying spatial differences in urban public transport services according to claim 5, characterized in that, The standardization of each indicator in the spatial service difference identification index system is performed, and the standardized indicator data is analyzed using the K-means clustering algorithm to identify and summarize various spatial regions, outputting the spatial difference identification results of urban public transport services, including: Z-score standardization was applied to the distance centrality index, the proportion of elderly people, the profile diversity index, the population balance index, the supply-demand ratio index, the net mobility intensity index, and the functional specialization index. K-means clustering algorithm was used to perform cluster analysis on the standardized index data; Based on the clustering results, various spatial regions are identified, including core service areas, balanced development areas, peripheral residential areas, and transitional mixed areas. Analyze the distribution of characteristic indicators of various spatial regions, and summarize the basic characteristics and service difference patterns of various spatial regions.

8. A device for identifying spatial differences in urban public transport services based on travel behavior data, characterized in that, The device includes: The acquisition and preprocessing module is used to acquire public transportation data and facility data of the target area, preprocess the public transportation data and facility data, and divide the target area into hexagonal grids as basic zoning units. The public transportation data includes public transportation card swipe data, public transportation station vector data, and public transportation route vector data. The public transportation card swipe data includes user identifier, travel group identifier, travel time, origin station, destination station, and travel route. The public transportation station vector data includes station name, station coordinates, and the route to which it belongs. The public transportation route vector data includes route name and route station sequence. The facility data includes land use vector data and POI data. The determination module is used to comprehensively determine the purpose of travel based on data of the origin station, destination station and surrounding facilities, combined with travel time and distance. The first construction module is used to build a four-dimensional travel profile tag system based on travel groups, travel purposes, travel time periods and travel distances, and generate a user-profile frequency matrix; The extraction module is used to normalize the user-profile frequency matrix, perform multidimensional extended analysis using singular value decomposition, and extract profile pattern features. The calculation module is used to map travel records to a hexagonal grid, aggregate and calculate the profile intensity index of each grid, and distinguish between demand-oriented profiles and supply-oriented profiles based on the travel purpose attributes in the profiles, and calculate the supply-demand ratio index of each grid. The second construction module is used to calculate the distance centrality index and the net flow intensity index, and combine them with the profile intensity index and the supply-demand ratio index to construct a spatial service difference identification index system. The identification module is used to standardize the various indicators in the spatial service difference identification indicator system, and uses the K-means clustering algorithm to analyze the standardized indicator data, identify and summarize various spatial regions, and output the spatial difference identification results of urban public transport services.

9. A computer device comprising a processor and a memory for storing a processor-executable program, characterized in that, When the processor executes the program stored in the memory, it implements the urban public transport service spatial difference identification method according to any one of claims 1-7.

10. A storage medium storing a program, characterized in that, When the program is executed by the processor, it implements the urban public transport service spatial difference identification method according to any one of claims 1-7.

Images