A method for identifying urban spatial organization patterns of multi-dimensional traffic flows by integrating map data

Abstract

The present invention provides a method for identifying the spatial organization pattern of a multivariate traffic flow city by integrating map data, comprising the following steps: step S1: obtaining railway, highway and air passenger train data and establishing a relationship matrix; step S2: setting the travel mode and obtaining the time distance of intercity travel; step S3: using the ratio of the number of trains to the time distance as a weight, and comprehensively weighting the railway, highway and air networks to construct a multivariate traffic flow network; step S4: constructing a multivariate traffic flow network community structure detection model based on K-shortest path and hierarchical clustering; step S5: comprehensively integrating the geographic network and the topological network for the community structure, identifying the spatial organization pattern, and deeply exploring the potential correlation structure of the multivariate traffic flow city network. The present invention effectively identifies the spatial organization pattern of the urban network by combining the characteristics that actually affect the traveler's experience and choice, such as the real-time road conditions of the highway, the travel cost of the train or flight, etc.

Description

Technical Field

[0001] The present invention relates to the technical field of urban network regional effect mining and analysis, in particular to a method for identifying urban spatial organization patterns of multi-traffic flows integrating map data. Background Art

[0002] As one of the basic public services, transportation infrastructure explores the multi-center, networked urban network space development model from the perspective of complex networks and individual travel behaviors. The rationality of its spatial distribution is directly related to the quality of life of residents. A community refers to a network subset composed of a cluster of closely connected nodes. Its characteristics are that the nodes within the community are closely connected, while the connections between nodes between communities are relatively sparse. Therefore, it is impossible to evaluate whether the transportation infrastructure and services within the urban agglomeration are well integrated. Summary of the invention

[0003] In view of this, the purpose of the present invention is to provide a method for identifying urban spatial organization patterns of multiple traffic flows that integrates map data. The present invention combines the characteristics that actually affect travelers' experience and choices, such as real-time road conditions, travel costs of trains or flights, to effectively identify urban network spatial organization patterns.

[0004] To achieve the above object, the present invention adopts the following technical solution: a multi-traffic flow urban spatial organization pattern recognition method integrating map data, comprising the following steps:

[0005] Step S1: Obtain railway, highway and air passenger transport data, classify multiple stations in the same city as a city node, and record the number of connections between stations to establish a relationship matrix;

[0006] Step S2: construct a multi-modal accessibility model based on the map API path planning interface, set the travel mode, and obtain the inter-city OD travel time and distance;

[0007] Step S3: adopt the P model and L model of complex network, and take the ratio of the number of flights to the time distance as the weight, and comprehensively weight the railway, highway and aviation networks to construct a multivariate traffic flow network;

[0008] Step S4: construct a multivariate traffic flow network community structure detection model based on K-shortest paths and hierarchical clustering, select geographic modularity for improvement, and perform community structure segmentation on the multivariate traffic flow network;

[0009] Step S5: Comprehensively integrate the geographic network and topological network for community structure, identify spatial organization patterns, and deeply explore the potential correlation structure of the multi-transportation flow urban network.

[0010] In a preferred embodiment, the step S1 is specifically as follows:

[0011] Step S11: using a crawler program to crawl railway, highway and air passenger transport schedule data and the longitude and latitude coordinates of the city center;

[0012] Step S12: multiple sites located in the same city are classified as one city node, and the number of connections between the sites is recorded;

[0013] Step S13: Using ArcGIS spatial connection tools, the coordinates of each city center point are mapped to each administrative division, the number of station shifts within each administrative division is summed up, and a relationship matrix is ​​established; the establishment of the relationship matrix is ​​as follows:

[0014]

[0015] In the formula, R ij is the traffic flow intensity from the starting point city i to the ending point city j, and R is the traffic flow intensity of the city.

[0016] In a preferred embodiment, the step S2 is specifically as follows:

[0017] Step S21: calling the map api route planning interface, taking the center point of the starting point city i as the starting point and the center point of the ending point city j as the ending point, simulating the real resident travel scenario;

[0018] Step S22: Establish a full connection between the indexes of the starting point city i and the ending point city j, and obtain the inter-city OD travel time distance by setting the traffic mode of the interface.

[0019] In a preferred embodiment, the inter-city OD travel time distance is obtained as follows:

[0020] Use the route planning interface in the map API to obtain the travel time and distance between the city center points OD, and consider multiple transportation modes and transfer combinations based on real-time traffic conditions; including the total travel time under the three travel modes of road, rail and air;

[0021] t (T,C,A) =t os +t transfer +t ss +t sd

[0022] Where: T, C and A represent the three modes of transportation: train, car and aviation respectively, t (T,C,A) represents the total travel time using three modes of transportation: train, car and air; t (T,C,A) It is divided into four stages: (1) The time t that residents take public transportation from the government center of city i to the nearby train station, bus station, and airport os ; (2) The time required to transfer between two stations ttransfer , including ticket collection, waiting time, and entry and exit time; (3) the running time between two stations t ss ; (4) Time t from the station to the center of city j by public transportation sd , and finally the total travel time between the two city centers is obtained.

[0023] In a preferred embodiment, the step S3 is specifically as follows:

[0024] Step S31: The time cost obtained in step S2 is used as the travel time distance in the model, and the P model is used to construct the railway network. The differences in travel costs of different trains need to be considered, and the final link weight is the sum of the weights of the high-speed trains and low-speed trains passing through it;

[0025] Step S32: Similarly, the ratio of the number of line connections to the travel time distance to the nth power is used as the link weight, and the L model is used to construct the highway network and the aviation network respectively;

[0026] Step S33: According to the passenger volume ratio reported by the transportation system, the weight of each transportation system is set, and the railway network, highway network and aviation network are comprehensively weighted to construct a comprehensive transportation network.

[0027] In a preferred embodiment, the construction of the comprehensive transportation network is as follows:

[0028] Step S331: The ratio of the number of trains m to the travel time distance to the power of n is regarded as the node edge weight, and the final link weight W between the two cities is T The high-speed train W G With low-speed train W K The weight and, similarly, directly construct the highway network W B and Aviation Network W A The formula is as follows:

[0029]

[0030]

[0031] In the formula, m G Indicates the number of high-speed trains, m K Indicates the number of low-speed trains, m B Indicates the number of highway bus trips, m A Indicates the number of air flights; Indicates the travel time and distance of high-speed trains obtained through the map API. Indicates the travel time and distance of low-speed trains obtained through the map API. Indicates the travel time and distance of the road obtained through the map API. Represents the air travel time distance obtained through the map API; n is the distance friction coefficient in the gravity model set to 2;

[0032] Step S332: Based on the three transportation networks, the railway network, highway network and aviation network are comprehensively weighted to construct a comprehensive transportation network. The city node edge link weight W C ; According to the proportion of passenger volume data of different transportation systems in the transportation bulletin, the weight of each transportation system is fitted. The proportion of passenger volume of road, railway and air transportation is 16.5:5.27:1 respectively. The formula is as follows:

[0033] W C =μW B +νW T +W A

[0034] Where: the μ value is uniformly set to 16.5, and the ν value is 5.27.

[0035] In a preferred embodiment, the step S4 is specifically as follows:

[0036] Step S41: using the YEN algorithm K-shortest path to quantify the proximity index between cities;

[0037] Step S42: According to the defined proximity index, a hierarchical clustering method is used to obtain a clustering tree to reveal the hierarchical structure of the network;

[0038] Step S42: In the clustering process, geographic modularity is introduced to perform community structure segmentation on the comprehensive weighted multivariate traffic flow network.

[0039] In a preferred embodiment, the proximity index between the quantified cities is specifically as follows:

[0040] Step S411: When quantizing the proximity index, use the weight W C The reciprocal of is the connection strength between two city nodes, and the proximity index between city i and city j is It is defined as the weighted combination of the k-th shortest paths, and its formula is as follows:

[0041]

[0042] Where: k and L k are the weight and length of the kth shortest path respectively;

[0043] Step S412: Next, set the K value and ω K value, in order to obtain the K value and the kth shortest path weight ω K, where for K values ​​from 1 to 3 with a step size of 1, ω K Experiment from 0 to 1 with a step size of 0.01 and select the final parameters.

[0044] In a preferred embodiment, the hierarchical clustering method is specifically as follows:

[0045] Step S421: hierarchical clustering is used to iteratively merge nodes according to the defined proximity index, and the proximity index between clusters α and β is updated as shown in the formula:

[0046]

[0047] Where: n α,β is the number of node pairs between clusters α and β; is the weighted combination of the k-th shortest paths;

[0048] Step S422: In each iteration, hierarchical clustering merges the cluster pairs with the smallest proximity index into a single cluster, and then updates the proximity index between this new cluster and all remaining clusters; finally, a dendrogram is formed to represent the bottom-up hierarchical structure of node clusters. When clusters α and β are merged into α+β, the proximity index between clusters α+β and γ is calculated as follows:

[0049]

[0050] The hierarchical clustering method obtains a tree structure. The defined weight is input into the geographic modularity, which defines the edge weight as the nth power of the distance between nodes. The geographic modularity is used to evaluate the quality of cluster detection and as an optimization condition for hierarchical clustering, as shown in the formula:

[0051]

[0052] In the formula, α is the community α, μ and ν are the nodes in the community α, is the community weight matrix, is the sum of the elements in each row of the matrix, w uv is the edge weight between nodes μ and ν, W sum is the weighted sum of the edges in the network;

[0053] In each iteration of the hierarchical clustering algorithm, the geographic modularity of each community is calculated. Taking communities C1 and C2 as an example, if Then communities C1 and C2 are merged, otherwise, communities C1 and C2 are marked as stop points. After constructing the cluster dendrogram, each branch of the dendrogram is traced back from the top layer until a cluster marked with a stop point is encountered, and it is interrupted at the stop point to form two communities. Each community will include a series of nodes; finally, the nodes on the traffic network are divided into a community set {α1, α2, ..., α n}, and ensure the overall geographic modularity The global maximum value can be obtained, that is, integrating geographic modularity into the hierarchical clustering algorithm can ensure the comprehensive weighted multivariate traffic flow network for community structure segmentation; among them, is the geographic modularity of each community, i is the community number, i=1,2....n.

[0054] In a preferred embodiment, the step S5 is specifically as follows:

[0055] Step S51: taking the community structure segmented in step S4 as the research unit, counting the shift connection strength between nodes;

[0056] Step S52: Based on the community structure segmentation results, the geographic network and topological network are integrated for visualization to identify the spatial organization pattern.

[0057] Compared with the prior art, the present invention has the following beneficial effects:

[0058] 1. The present invention uses Baidu Map API to obtain real-time traffic cost data and explores the spatial organization pattern of multi-dimensional traffic flow networks from the perspective of actual travel behavior;

[0059] 2. The present invention organically combines the Internet open map with the community discovery algorithm, uses the ratio of the number of flights to the travel time distance as the weight to quantify the proximity between cities, and selects the geographic modularity for improvement, and constructs a multivariate traffic flow network community structure detection model based on K-shortest path and hierarchical clustering;

[0060] 3. Based on the community structure research unit, the present invention constructs a comprehensive integrated geographic network and topological network to visualize the spatial organization pattern. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 is a schematic diagram of a community structure detection process of a method according to a preferred embodiment of the present invention;

[0062] Figure 2 The spatial distribution of the multi-dimensional traffic flow network in the Yangtze River Economic Belt in the preferred embodiment of the present invention;

[0063] Figure 3 The multi-transportation mode accessibility model in the preferred embodiment of the present invention;

[0064] Figure 4 A schematic diagram of the community structure detection process and results, i.e., the community clustering process on a multivariate traffic flow network, in a preferred embodiment of the present invention;

[0065] Figure 5 The community structure detection process and results in the preferred embodiment of the present invention are a community structure diagram based on a multi-element traffic flow network;

[0066] Figure 6 In the preferred embodiment of the present invention, the spatial distribution of the single core spatial organization pattern;

[0067] Figure 7 The spatial distribution of the dual-core spatial organization pattern in the preferred embodiment of the present invention;

[0068] Figure 8 In a preferred embodiment of the present invention, the multi-core spatial organization pattern is spatially distributed. DETAILED DESCRIPTION

[0069] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0070] It should be noted that the following detailed descriptions are illustrative and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0071] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application; as used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or their combinations.

[0072] Please refer to Figures 1 to 8 The present invention provides a method for identifying a multi-dimensional traffic flow urban network spatial organization pattern by integrating map data, which specifically comprises the following steps:

[0073] Step S1: Use Python language to crawl railway, highway and air passenger flight data, classify multiple stations in the same city as a city node, and record the number of connections between stations to establish a relationship matrix;

[0074] Step S2: Build a multi-modal accessibility model based on the Baidu Map API path planning interface, set the travel mode, and obtain the inter-city travel OD time distance.

[0075] Step S3: Using the P model and L model of complex networks, taking the ratio of the number of flights to the time distance as the weight, and comprehensively weighting the railway, highway and aviation networks to construct a multivariate traffic flow network;

[0076] Step S4: construct a multivariate traffic flow network community structure detection model based on K-shortest paths and hierarchical clustering, select geographic modularity for improvement, and divide the community structure of the multivariate traffic flow network;

[0077] Step S5: Comprehensively integrate the geographic network and topological network for community structure, identify spatial organization patterns, and deeply explore the potential correlation structure of the urban network based on multi-dimensional traffic flow.

[0078] In this example, the step S1 uses Python to crawl railway, highway and air passenger schedule data, classifies multiple sites in the same city as a city node, and records the number of contacts between sites to establish a relationship matrix. Specifically, the following steps are included:

[0079] Step S11: Use Python to write a crawler program to crawl railway, highway and air passenger schedule data from Ctrip.com (https: / / www.ctrip.com). In order to ensure data integrity as much as possible, crawl Qunar.com (https: / / www.qunar.com) and Railway Network (https: / / www.12306.cn / ) at the same time, and merge the results to supplement the data. Correct the data through data cleaning, manual inspection, etc. to ensure the accuracy of the data.

[0080] Step S12: multiple sites located in the same city are classified as one city node, and the number of connections between the sites is recorded.

[0081] Step S13: Using ArcGIS spatial connection tools, the coordinates of each city center point are mapped to each administrative division. For each administrative division, the number of stations and shifts within it is summed up, and a relationship matrix is ​​established to obtain a multivariate traffic flow spatial distribution pattern used to describe the spatial correlation intensity of the urban network. The formula is as follows:

[0082]

[0083] In the formula, R ij is the traffic flow intensity from the starting city i to the ending city j.

[0084] In this example, in step S2, the intercity OD travel time and distance based on multiple transportation modes are obtained by calling the Baidu map API path planning interface.

[0085] Step S21: Write a program in Python to convert the coordinates into the Baidu coordinate system, and then call the Baidu map api path planning interface, using the center point of the starting city i as the starting point and the center point of the ending city j as the ending point, simulating the real resident travel scenario.

[0086] Step S23: Establish a full connection based on the indexes of the center point of the starting point city i and the center point of the ending point city j, and obtain the inter-city OD travel time distance by setting the traffic mode of the interface.

[0087] In this implementation, the multi-mode time cost returned by the Baidu map API path planning interface in step S2 is used as the time distance in the community structure detection model, as follows:

[0088] Using the route planning interface in the map API, we can obtain the travel time and distance between the city center points OD, and based on the real-time traffic conditions, consider multiple modes of transportation and transfer combinations to achieve a "door-to-door" travel method. This includes the total travel time under the three travel modes of road, rail and air.

[0089] t (T,C,A) =t os +t transfer +t ss +t sd

[0090] Where: T, C and A represent the three modes of transportation: train, car and aviation respectively, t (T,C,A) represents the total travel time using three modes of transportation: train, car and air, t (T,C,A) It is divided into four stages: (1) The time t that residents take public transportation from the government center of city i to the nearby train station, bus station, and airport os ; (2) The time required to transfer between two stations t transfer , including ticket collection, waiting time, and entry and exit time; (3) the running time between two stations t ss ; (4) Time t from the station to the center of city j by public transportation sd , and finally the total travel time between the two city centers is obtained.

[0091] In the implementation of this example, the P model and L model of the complex network are used in step S3, the ratio of the number of trains to the time distance is used as the weight, and the railway, highway and aviation networks are comprehensively weighted to construct a multi-dimensional traffic flow network.

[0092] Step S31: The time cost obtained in step S2 is used as the travel time distance in the model, and the P model is used to construct the railway network. The differences in travel costs of different trains need to be considered, and the final link weight is the sum of the weights of the high-speed trains and low-speed trains passing through it;

[0093] Step S32: Similarly, the ratio of the number of line connections to the travel time distance n is used as the link weight, and the L model is used to construct the road network and the aviation network respectively.

[0094] Step S33: According to the passenger volume ratio reported by the transportation system, the weight of each transportation system is set, and the basic idea of ​​comprehensively weighting the railway network, highway network and aviation network to construct a comprehensive transportation network is:

[0095] (1): The ratio of the number of trains (m) to the nth power of the time distance is regarded as the node edge weight, and the final link weight between the two cities is W T The high-speed train W G With low-speed train W K The weight and, similarly, directly construct the highway network (W B ) and aviation network (W A )The formula is as follows:

[0096]

[0097]

[0098] In the formula, m G Indicates the number of high-speed trains, m K Indicates the number of low-speed trains, m B Indicates the number of highway bus trips; Indicates the travel time and distance of high-speed trains obtained through the map API. Indicates the travel time and distance of low-speed trains obtained through the map API. Indicates the travel time and distance of the road obtained through the map API. Represents the aviation travel time distance obtained through the map API; n is the distance friction coefficient in the gravity model is set to 2.

[0099] (2): Based on the above three types of transportation networks, the railway network, highway network and aviation network are comprehensively weighted to construct a comprehensive transportation network. The edge link weight of the city node is W C According to the proportion of passenger volume data of different transportation systems in the transportation bulletin, the weight of each transportation system is fitted. The proportion of passenger volume of highway, railway and air transportation is 16.5:5.27:1 respectively. The formula is as follows:

[0100] W C =μW B +νW T +W A

[0101] Where: the μ value is uniformly set to 16.5, and the ν value is 5.27.

[0102] In this implementation, in step S4, a multivariate traffic flow network community structure detection model based on K-shortest paths and hierarchical clustering is constructed, and geographic modularity is selected for improvement to perform community structure segmentation on the multivariate traffic flow network. Specifically, the following steps are included:

[0103] Step S41: using the YEN algorithm K-shortest path to quantify the proximity index between cities;

[0104] (1): When quantifying the proximity index, the weight W C The reciprocal of is the connection strength between two city nodes, and the proximity index between node i and node j is It is defined as the weighted combination of the k-th shortest paths, and its formula is as follows:

[0105]

[0106] Where: k and L k are the weight and length of the kth shortest path respectively;

[0107] (2): Next, set the K value and ω K In order to obtain the appropriate K value and the kth shortest path weight ω K , where for K values ​​from 1 to 3 with a step size of 1, ω K A series of experiments were conducted from 0 to 1 with a step size of 0.01. When K = 2, ω1 = 0.98, ω2 = 0.02, the geographic modularity (Q = 0.728) of the multivariate traffic flow network reached its maximum value.

[0108] Step S42: According to the defined proximity index, further reveal the merging process of these city cluster nodes, and use the hierarchical clustering method to obtain a clustering tree to reveal the hierarchical structure of the network. Figure 4 , taking the Yangtze River Delta, the middle reaches of the Yangtze River and the three major urban agglomerations of Chengdu and Chongqing as the research objects, some significant inflection points were selected, among which the community proximity indexes of these inflection points were 0.86, 3.16, 10.46, 19.95 and 41.16, and the corresponding numbers of communities on the network were 120, 89, 52, 34 and 21 respectively. The specific steps include:

[0109] (1): According to the proximity index defined above, hierarchical clustering is used to iteratively merge nodes, and the proximity index between clusters α and β is updated as shown in the formula:

[0110]

[0111] Where: nα,β is the number of node pairs between clusters α and β; is the weighted combination of K shortest paths.

[0112] (2): In each iteration, hierarchical clustering merges the cluster pairs with the smallest proximity index into a single cluster, and then updates the proximity index between this new cluster and all remaining clusters. Finally, a dendrogram is formed to represent the bottom-up hierarchical structure of node clusters. When clusters α and β are merged into α+β, the proximity index between clusters α+β and γ is calculated as follows:

[0113]

[0114] Step S43: In the clustering process, the geographic modularity is introduced to segment the community structure for the comprehensive weighted multivariate traffic flow network. Figure 5 , using the constructed TCD community structure detection model, we found that there are 24 communities in the Yangtze River Economic Belt, among which 6 communities have relatively close internal connections, which are called "urban economic zones", and are named Guizhou Community, Yunnan Community, Anhui Community, Sichuan and Chongqing Community, Hubei, Hunan and Jiangxi Community and Jiangsu, Zhejiang and Shanghai Community respectively.

[0115] The hierarchical clustering method above obtains a tree structure. In order to quantify the impact of spatial distance on connection strength, the defined weight is input into the geo-modularity, which defines the edge weight as the nth power of the distance between nodes. The geo-modularity is used to evaluate the quality of cluster detection and as an optimization condition for hierarchical clustering, as shown in the formula:

[0116]

[0117] In the formula, α is the community α, μ and ν are the nodes in the community α, is the community weight matrix, is the sum of the elements in each row of the matrix, w uv is the edge weight between nodes μ and ν, W sum is the weighted sum of the edges in the network;

[0118] In each iteration of the hierarchical clustering algorithm, the geographic modularity of each community is calculated. Taking communities C1 and C2 as an example, if Then communities C1 and C2 are merged, otherwise, communities C1 and C2 are marked as "stop points". After the clustering dendrogram is constructed, each branch of the dendrogram is traced back from the top layer until a cluster marked with a "stop point" is encountered, and it is interrupted at the stop point to form two communities, each of which will include a series of nodes. Finally, the nodes on the transportation network are divided into a community set {α1, α2, ..., α n}, and ensure the overall geographic modularity The global maximum value can be obtained, that is, integrating geographic modularity into the hierarchical clustering algorithm can ensure the optimal community structure segmentation of the comprehensive weighted multivariate traffic flow network; is the geographic modularity of each community, i is the community number, i=1,2....n.

[0119] In this embodiment, the step S5 integrates the geographic network and the topological network, identifies the spatial organization pattern, and deeply explores the potential correlation structure of the multi-transportation urban network. The step S5 specifically includes the following steps:

[0120] Step S51: Taking the community structure segmented in step S4 as the research unit, the shift connection strength between nodes is counted to construct a community structure subnetwork.

[0121] Step S52: Visualize the segmentation results of the community structure detection model. Referring to Figures 6, 7, and 8, the regional spatial organization pattern based on the perspective of multiple traffic flow connections can be roughly divided into three development patterns: single-core, dual-core, and multi-core. Among them, the three communities with Guizhou, Yunnan, and Anhui as the cores present a single-core development pattern; the Sichuan-Chongqing community presents a dual-core development pattern; the Hubei-Hunan-Jiangxi community and the Jiangsu-Zhejiang-Shanghai community present a multi-center development pattern.

[0122] In this embodiment, a multi-dimensional traffic flow urban network spatial organization pattern recognition system integrating map data is also provided, including data collection and network construction, community structure detection model construction and spatial organization pattern recognition;

[0123] Data collection and network construction use Python language to crawl and obtain railway, highway and air passenger flight data, classify multiple stations in the same city as a city node, and record the number of connections between stations to establish a relationship matrix; build a multimodal accessibility model based on the Baidu Map API path planning interface, set the travel mode, and obtain the OD travel time and distance.

[0124] Construct a community structure detection model and identify spatial organization patterns. Construct a multi-transportation network community structure detection model based on K-shortest path and hierarchical clustering, select geographic modularity for improvement, and perform community structure segmentation on the multi-transportation network. Take community structure as the research unit, comprehensively integrate geographic network and topological network, identify spatial organization patterns, and deeply explore the potential correlation structure of the urban network based on multi-transportation flow.

[0125] The above description is only a preferred embodiment of the present invention. All equivalent changes and modifications made according to the scope of the patent application of the present invention should fall within the scope of the present invention.

Claims

1. A method for identifying urban spatial organization patterns of multi-traffic flows by integrating map data, characterized in that: The following steps are involved: Step S1: Obtain railway, highway and air passenger transport data, classify multiple stations in the same city as a city node, and record the number of connections between stations to establish a relationship matrix; Step S2: Set the travel mode based on the map API route planning interface and obtain the inter-city OD travel time and distance; Step S3: adopt the P model and L model of complex network, and take the ratio of the number of flights to the time distance as the weight, and comprehensively weight the railway, highway and aviation networks to construct a multivariate traffic flow network; Step S4: construct a multivariate traffic flow network community structure detection model based on K-shortest paths and hierarchical clustering, select geographic modularity for improvement, and perform community structure segmentation on the multivariate traffic flow network; Step S5: Comprehensively integrate the geographic network and topological network for the community structure, identify the spatial organization pattern, and deeply explore the potential correlation structure of the multi-transportation flow urban network; The step S3 is specifically as follows: Step S31: The time cost obtained in step S2 is used as the travel time distance in the model, and the P model is used to construct the railway network. The differences in travel costs of different trains need to be considered, and the final link weight is the sum of the weights of the high-speed trains and low-speed trains passing through it; Step S32: Similarly, the ratio of the number of line connections to the travel time distance to the nth power is used as the link weight, and the L model is used to construct the highway network and the aviation network respectively; Step S33: according to the passenger volume ratio reported by the transportation system, the weight of each transportation system is set, and the railway network, highway network and aviation network are comprehensively weighted to construct a comprehensive transportation network; The construction of the comprehensive transportation network is as follows: Step S331: The ratio of the number of trains m to the travel time distance to the power of n is regarded as the node edge weight, and the final link weight W between the two cities is T The high-speed train W G With low-speed train W K The weight and, similarly, directly construct the highway network W B and Aviation Network W A The formula is as follows: Where: m G Indicates the number of high-speed trains, m K Indicates the number of low-speed trains, m B Indicates the number of highway bus trips, m A Indicates the number of air flights; Indicates the travel time and distance of high-speed trains obtained through the map API. Indicates the travel time and distance of low-speed trains obtained through the map API. Indicates the travel time and distance of the road obtained through the map API. Represents the aviation travel time distance obtained through the map API; n is the distance friction coefficient in the gravity model; Step S332: Based on the three transportation networks, the railway network, highway network and aviation network are comprehensively weighted to construct a comprehensive transportation network. The city node edge link weight W C ; According to the proportion of passenger volume data of different transportation systems in the transportation bulletin, the weight of each transportation system is fitted. The proportion of passenger volume of road, railway and air transportation is 16.5:5.27:1 respectively. The formula is as follows: IN C =μW B +νW T +W A Where: μ is uniformly set to 16.5, ν is 5.27; The step S4 is specifically as follows: Step S41: using the YEN algorithm K-shortest path to quantify the proximity index between cities; Step S42: According to the defined proximity index, a hierarchical clustering method is used to obtain a clustering tree to reveal the hierarchical structure of the network; Step S42: In the clustering process, geographic modularity is introduced to perform community structure segmentation on the comprehensive weighted multivariate traffic flow network; The quantitative proximity indicators between cities are as follows: Step S411: When quantizing the proximity index, use the weight W C The reciprocal of is the connection strength between two city nodes, and the proximity index between city i and city j is It is defined as the weighted combination of the k-th shortest paths, and its formula is as follows: Among them, ω k and L k are the weight and length of the kth shortest path respectively; Step S412: Next, set the K value and ω K value, in order to obtain the K value and the kth shortest path weight ω K , where for K values ​​from 1 to 3 with a step size of 1, ω K Experiment from 0 to 1 with a step size of 0.01 to select the final parameters; The hierarchical clustering method is as follows: Step S421: hierarchical clustering is used to iteratively merge nodes according to the defined proximity index, and the proximity index between clusters α and β is updated as shown in the formula: Where: n α,β is the number of node pairs between clusters α and β; is the weighted combination of the k-th shortest paths; Step S422: In each iteration, hierarchical clustering merges the cluster pairs with the smallest proximity index into a single cluster, and then updates the proximity index between this new cluster and all remaining clusters; finally, a dendrogram is formed to represent the bottom-up hierarchical structure of node clusters. When clusters α and β are merged into α+β, the proximity index between clusters α+β and γ is calculated as follows: In the multivariate traffic flow network community structure detection model, the defined weight W is used. C W B W T ,,,W A The input is into the geographic modularity formula and applied to the above hierarchical clustering process to find the optimal community segmentation; geographic modularity is used to evaluate the quality of cluster detection and as an optimization condition for hierarchical clustering, as shown in the formula: In the formula, α is the community α, μ and ν are the nodes in the community α, is the community weight matrix, is the sum of the elements in each row of the matrix, w uv is the edge weight between nodes μ and ν, W sum is the weighted sum of the edges in the network; in each iteration of the hierarchical clustering algorithm, the geographic modularity of each community will be calculated, denoted by community C1 and community C2. If Then community C1 and community C2 are merged, otherwise, community C1 and community C2 are marked as stop points. After constructing the cluster dendrogram, each branch of the dendrogram is traced back from the top layer until a cluster marked with a stop point is encountered, and it is interrupted at the stop point to form two communities. Each community will include a series of nodes; finally, the nodes on the traffic network are divided into a community set {α1, α2, ..., α n }, and ensure the overall geographic modularity The global maximum value can be obtained, that is, integrating geographic modularity into the hierarchical clustering algorithm can ensure the comprehensive weighted multivariate traffic flow network for community structure segmentation; among them, is the geographic modularity of each community, i is the community number, i=1,2....n.

2. The method for identifying urban spatial organization patterns of multi-traffic flows by integrating map data according to claim 1 is characterized in that: The step S1 is specifically as follows: Step S11: using a crawler program to crawl railway, highway and air passenger transport schedule data and the longitude and latitude coordinates of the city center; Step S12: multiple sites located in the same city are classified as one city node, and the number of connections between the sites is recorded; Step S13: Using ArcGIS spatial connection tools, the coordinates of each city center point are mapped to each administrative division, the number of station shifts within each administrative division is summed up, and a relationship matrix is ​​established; the establishment of the relationship matrix is ​​as follows: In the formula, R ij is the traffic flow intensity from the starting point city i to the ending point city j, and R is the traffic flow intensity of the city.

3. The method for identifying urban spatial organization patterns of multi-traffic flows by integrating map data according to claim 1, characterized in that: The step S2 is specifically as follows: Step S21: calling the map api route planning interface, taking the center point of the starting point city i as the starting point and the center point of the ending point city j as the ending point, simulating the real resident travel scenario; Step S22: Establish a full connection between the indexes of the starting point city i and the ending point city j, and obtain the inter-city OD travel time distance by setting the traffic mode of the interface.

4. The method for identifying urban spatial organization patterns of multi-traffic flows by integrating map data according to claim 3 is characterized in that: The inter-city OD travel time and distance are obtained as follows: Use the route planning interface in the map API to obtain the travel time and distance between the city center points OD, and consider multiple transportation modes and transfer combinations based on real-time traffic conditions; including the total travel time under the three travel modes of road, rail and air; t (T,C,A) =t os +t transfer +t ss +t sd Where: T, C and A represent the three modes of transportation: train, car and aviation respectively, t (T,C,A) represents the total travel time using three modes of transportation: train, car and air; t (T,C,A) It is divided into four stages: (1) The time t that residents take public transportation from the government center of city i to the nearby train station, bus station, and airport os ; (2) The time required to transfer between two stations t transfer , including ticket collection, waiting time, and entry and exit time; (3) Running time between two sites t ss ; (4) Time t from the station to the center of city j by public transportation sd , and finally the total travel time between the two city centers is obtained.

5. The method for identifying urban spatial organization patterns of multi-traffic flows by integrating map data according to claim 1, characterized in that: The step S5 is specifically as follows: Step S51: taking the community structure segmented in step S4 as the research unit, counting the shift connection strength between nodes; Step S52: Based on the community structure segmentation results, the geographic network and topological network are integrated for visualization to identify the spatial organization pattern.

Images