Urban crowd activity and space structure dynamic deduction method and system based on single traffic flow, terminal and storage medium

Abstract

The invention belongs to the technical field of traffic geographic information analysis, and discloses an urban crowd activity and space structure dynamic deduction method and system based on single traffic flow, a terminal and a storage medium. Identifying a spatial distribution mode through a flow similarity measurement method in combination with a hierarchical clustering algorithm; carrying out quantity aggregation on the identified spatial distribution mode according to grid units, and constructing a spatial grade distribution model; combining the urban built environment features with spatial grade distribution in the spatial grade distribution model, and constructing a comprehensive data set suitable for machine learning regression modeling; and based on the comprehensive data set, in combination with a machine learning model interpretation method of a game theory, quantitatively analyzing a nonlinear influence mechanism of urban built environment characteristics on different spatial distribution modes, and deducing internal relevance between urban crowd activities and spatial structures. The dynamic evolution process of the urban space structure is comprehensively realized.

Description

Technical Field

[0001] This invention relates to the field of traffic geographic information analysis technology, and in particular to a method, system, terminal, and storage medium for dynamic extrapolation of urban population activity and spatial structure based on single-vehicle flow. Background Technology

[0002] With the acceleration of urbanization and the continuous increase in population density, urban transportation systems face structural problems such as increased commuting pressure, severe separation of work and residence, and unbalanced functional area layout. To effectively address these challenges and achieve efficient organization and sustainable development of urban space, it is necessary to conduct in-depth analysis of urban structure and its dynamic evolution using new technologies. Understanding urban travel behavior from the perspective of geographical mobility helps reveal the interaction patterns between different regions, and combining analysis with multiple mobility patterns can more comprehensively reflect the complex dynamic evolution mechanism of cities.

[0003] The development of big data and sensing technologies has led to an increasing abundance of high spatiotemporal resolution data on urban traffic flow. Shared bicycle travel data, with its clearly defined origin-destination (OD) points, timestamps, and wide coverage, has become an important data source for studying urban micro-level travel behavior. Existing research identifies urban travel hotspots through clustering and statistical analysis of shared bicycle OD data, further combining this with spatial analysis using population distribution, land use data, and built environment data. Other studies analyze urban morphology based on the spatiotemporal evolution characteristics of cycling flow, utilizing traditional statistical models such as factor analysis, principal component analysis, and maximum likelihood analysis to identify key driving factors. However, urban micro-level travel behavior is diverse, heterogeneous, selective, and dynamic; that is, the subjective intentions of human activity are difficult to predict or describe using simple physical models. Traditional models are insufficient in revealing the complex nonlinear relationships and threshold effects behind travel behavior; often, after conditions reach a certain critical point, changes in variables no longer have a significant impact on usage patterns. Furthermore, existing research lacks systematic analysis of multiple flow patterns, failing to comprehensively reflect the dynamic evolution of urban spatial structure and limiting in-depth research and scientific intervention in complex urban dynamic structures.

[0004] In summary, existing technologies still have the problem of failing to fully reflect the dynamic evolution of urban spatial structure, therefore, existing technologies need to be improved. Summary of the Invention

[0005] The technical problem to be solved by the present invention is that, in response to the deficiencies of the prior art, the present invention provides a method, system, terminal and storage medium for dynamic extrapolation of urban population activity and spatial structure based on single-vehicle flow, so as to solve the problem that the prior art is insufficient in systematic analysis of multiple flow patterns and is difficult to fully reflect the dynamic evolution process of urban spatial structure.

[0006] The technical solution adopted by this invention to solve the technical problem is as follows: In a first aspect, the present invention provides a method for dynamically extrapolating urban population activity and spatial structure based on single-vehicle flow, including: A cycling flow network is constructed based on shared bicycle order data within the city, and spatial distribution patterns are identified by combining flow similarity measurement methods with hierarchical clustering algorithms. The identified spatial distribution patterns are aggregated according to the number of grid cells to construct a spatial hierarchical distribution model; By combining the characteristics of the urban built environment with the spatial hierarchy distribution in the aforementioned spatial hierarchy distribution model, a comprehensive dataset suitable for machine learning regression modeling is constructed. Based on the comprehensive dataset, and combined with the machine learning model interpretation method of game theory, the nonlinear influence mechanism of the urban built environment characteristics on different spatial distribution patterns is quantitatively analyzed, and the intrinsic correlation between urban population activities and spatial structure is deduced.

[0007] In one implementation, the construction of the cycling flow network, based on shared bicycle order data within a city, identifies spatial distribution patterns using a flow similarity measurement method combined with a hierarchical clustering algorithm, including: Obtain shared bicycle order data within the city limits; The travel flow trajectory is obtained based on the shared bicycle order data; The cycling flow network is constructed based on the travel flow trajectory. Based on the cycling flow network, the similarity of cycling flows is evaluated using the flow similarity measurement method, and the spatial distribution pattern is identified using the hierarchical clustering algorithm.

[0008] In one implementation, the evaluation of cycling flow similarity using the flow similarity measurement method includes: For any two cycling paths in the cycling flow network, a dynamically adjustable spatial buffer zone is constructed with the starting point and ending point of each cycling flow path as the center. When the starting point and ending point of two cycling paths simultaneously fall into the spatial buffer zone, the two cycling paths are determined to be spatially similar flow paths.

[0009] In one implementation, the spatial distribution pattern includes: a clustering pattern, a diverging pattern, and a converging pattern; The method of identifying the spatial distribution pattern using the hierarchical clustering algorithm includes: Each of the aforementioned cycling flow paths is initialized as an independent flow cluster; Calculate the midpoint position of each cycling flow path, construct a KD tree spatial index structure based on all midpoint positions, and merge flow clusters from bottom to top based on the KD tree spatial index structure to identify the clustering pattern; Based on the starting point proximity discrimination method, a set of cycling flow paths with adjacent starting points and random directions is selected, and the divergence pattern is identified. Based on the endpoint proximity discrimination method, a set of cycling flow paths with adjacent endpoints and similar directions is selected, and the convergence pattern is identified.

[0010] In one implementation, the step of aggregating the identified spatial distribution patterns according to the number of grid cells to construct a spatial hierarchy distribution model includes: Using a gridded spatial statistical method, the total number of starting points and ending points of cycling flows within the grid under the clustering mode is counted, the number of starting points under the diverging mode is counted, and the number of ending points under the converging mode is counted to obtain the scale level of each mode. The spatial hierarchy distribution model is constructed based on the statistically analyzed scale levels.

[0011] In one implementation, the step of combining urban built environment characteristics with the spatial hierarchy distribution in the spatial hierarchy distribution model to construct a comprehensive dataset suitable for machine learning regression modeling includes: A dataset containing urban built environment characteristics is constructed and correlated with the spatial levels in the spatial level distribution model to obtain the comprehensive dataset; wherein, the urban built environment characteristics include: urban density characteristics, urban design characteristics, urban diversity characteristics, public transportation accessibility characteristics, and destination accessibility characteristics.

[0012] In one implementation, based on the comprehensive dataset and combined with game theory-based machine learning models, the nonlinear influence mechanism of the urban built environment characteristics on different spatial distribution patterns is quantitatively analyzed to deduce the intrinsic correlation between urban population activities and spatial structure, including: Based on the comprehensive dataset, a regression model is established using a machine learning algorithm with nonlinear fitting capabilities; Based on the regression model, the contribution of the urban built environment characteristics to the prediction of different spatial distribution patterns is quantitatively analyzed using a game theory-based machine learning model interpretation method. Based on the predicted contribution, the global feature importance ranking and local feature dependency map of the urban built environment features are output, and the inference results of the intrinsic correlation between urban population activities and spatial structure are obtained.

[0013] Secondly, the present invention provides a dynamic simulation system for urban population activity and spatial structure based on single-vehicle flow, comprising: The spatial distribution pattern recognition module is used to construct a cycling flow network. Based on shared bicycle order data within the city, it identifies spatial distribution patterns by combining flow similarity measurement methods with hierarchical clustering algorithms. The spatial hierarchy distribution model construction module is used to aggregate the identified spatial distribution patterns according to the number of grid cells to construct a spatial hierarchy distribution model. The comprehensive dataset construction module is used to combine the characteristics of the urban built environment with the spatial hierarchy distribution in the spatial hierarchy distribution model to construct a comprehensive dataset suitable for machine learning regression modeling. The cycling flow pattern interpretation module is used to quantitatively analyze the nonlinear influence mechanism of the urban built environment characteristics on different spatial distribution patterns based on the comprehensive dataset and combined with the machine learning model interpretation method of game theory, and to deduce the intrinsic correlation between urban population activities and spatial structure.

[0014] Thirdly, the present invention provides a terminal, comprising: a processor and a memory, wherein the memory stores a dynamic simulation program for urban population activity and spatial structure based on single-vehicle flow, and the dynamic simulation program for urban population activity and spatial structure based on single-vehicle flow is executed by the processor to implement the operation of the dynamic simulation method for urban population activity and spatial structure based on single-vehicle flow as described in the first aspect.

[0015] Fourthly, the present invention also provides a computer-readable storage medium storing a dynamic simulation program for urban population activity and spatial structure based on single-vehicle flow. When executed by a processor, the dynamic simulation program for urban population activity and spatial structure based on single-vehicle flow is used to implement the operation of the dynamic simulation method for urban population activity and spatial structure based on single-vehicle flow as described in the first aspect.

[0016] The present invention, by employing the above technical solution, has the following effects: 1) This invention systematically models and spatially analyzes the travel flow of shared bicycles. Based on the spatial distribution characteristics and the interaction between the starting points, the flow patterns of shared bicycles can be divided into clustered, divergent, and convergent types. An interpretable machine learning model is introduced to quantify the nonlinear influence of urban built environment characteristics on different flow patterns.

[0017] 2) This invention can not only reveal the characteristics of the multi-center structure and functional distribution of cities reflected by different types of flow patterns, but also realize the dynamic deduction of urban spatial structure based on the spatiotemporal evolution of single-vehicle travel flow, depict the role transformation and spatial connection changes of regions at different times, thereby showing the evolution law of urban spatial pattern.

[0018] 3) This invention can overcome the shortcomings of existing technologies in pattern recognition and formation mechanism explanation, and provide a basis for predicting the evolution trend of urban functional areas, optimizing the layout of the built environment, promoting sustainable travel, and promoting balanced urban spatial development. Attached Figure Description

[0019] 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.

[0020] Figure 1 This is a flowchart of the method for dynamically extrapolating urban population activities and spatial structure based on single-vehicle flow in this invention.

[0021] Figure 2 This is a technical flowchart of the present invention.

[0022] Figure 3 This is a schematic diagram of the shared bicycle flow similarity measurement in this invention.

[0023] Figure 4 This is a schematic diagram of pattern recognition and spatial aggregation in this invention.

[0024] Figure 5 This is a functional schematic diagram of the terminal in one implementation of the present invention.

[0025] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0027] Exemplary methods The development of big data and sensing technologies has led to an increasing abundance of high spatiotemporal resolution data on urban traffic flow. Shared bicycle travel data, with its clearly defined origin-destination (OD) points, timestamps, and wide coverage, has become an important data source for studying urban micro-level travel behavior. Existing research identifies urban travel hotspots through clustering and statistical analysis of shared bicycle OD data, further combining this with spatial analysis using population distribution, land use data, and built environment data. Other studies analyze urban morphology based on the spatiotemporal evolution characteristics of cycling flow, utilizing traditional statistical models such as factor analysis, principal component analysis, and maximum likelihood analysis to identify key driving factors. However, urban micro-level travel behavior is diverse, heterogeneous, selective, and dynamic; that is, the subjective intentions of human activity are difficult to predict or describe using simple physical models. Traditional models are insufficient in revealing the complex nonlinear relationships and threshold effects behind travel behavior; often, after conditions reach a certain critical point, changes in variables no longer have a significant impact on usage patterns. Furthermore, existing research lacks systematic analysis of multiple flow patterns, failing to comprehensively reflect the dynamic evolution of urban spatial structure and limiting in-depth research and scientific intervention in complex urban dynamic structures.

[0028] To address the above-mentioned technical problems, this invention provides a method for dynamically extrapolating urban population activities and spatial structure based on bicycle flow. The method includes: constructing a cycling flow network; identifying spatial distribution patterns based on shared bicycle order data within the city using a flow similarity measurement method combined with a hierarchical clustering algorithm; aggregating the identified spatial distribution patterns according to grid units to construct a spatial hierarchy distribution model; combining urban built environment characteristics with the spatial hierarchy distribution in the spatial hierarchy distribution model to construct a comprehensive dataset suitable for machine learning regression modeling; and based on the comprehensive dataset, using game theory-based machine learning model interpretation methods to quantitatively analyze the nonlinear influence mechanism of urban built environment characteristics on different spatial distribution patterns, and extrapolate the intrinsic correlation between urban population activities and spatial structure. This invention comprehensively realizes the dynamic evolution process of urban spatial structure.

[0029] like Figure 1 As shown, this embodiment of the invention provides a method for dynamically extrapolating urban population activity and spatial structure based on single-vehicle flow, including the following steps: Step S100: Construct a cycling flow network. Based on shared bicycle order data within the city, identify spatial distribution patterns by combining flow similarity measurement methods with hierarchical clustering algorithms.

[0030] In this embodiment, from the perspective of "mobility space", the shared bicycle travel flow data is systematically modeled and spatially analyzed. Based on the spatial distribution characteristics and the interaction relationship between the starting points, the flow patterns of shared bicycles are divided into clustering, diverging, and converging types. An interpretable machine learning model combining the SHAP method (a model interpretability method based on game theory Shapley value, used to quantify the contribution of each feature in the machine learning model to the prediction result) is introduced to quantify the nonlinear influence of urban built environment characteristics on different flow patterns.

[0031] The method provided in this embodiment not only reveals the urban polycentric structure and functional distribution characteristics reflected by different types of flow patterns, but also enables dynamic deduction of urban spatial structure based on the spatiotemporal evolution of vehicular traffic flow, characterizing the role transformation and spatial connection changes of regions at different times, thereby demonstrating the evolutionary laws of urban spatial patterns. This method overcomes the shortcomings of existing technologies in pattern recognition and formation mechanism explanation, establishes a systematic framework for urban dynamic structure analysis, and provides scientific basis and technical support for predicting the evolution trend of urban functional areas, optimizing the built environment layout, promoting sustainable mobility, and driving balanced urban spatial development.

[0032] The technical process of the method provided in this embodiment is as follows: Figure 2 As shown, it mainly consists of four core modules: cycling flow construction, spatial pattern recognition, built environment modeling, and interpretation and analysis. First, a refined cycling flow network is constructed. For shared bicycle order data within the city, a hierarchical clustering algorithm based on KD-tree (a tree-like data structure that stores instance points in K-dimensional space for fast retrieval) is used to achieve high-precision identification of spatial distribution patterns such as clustering, divergence, and convergence. This breaks through the limitation of existing research that only analyzes single patterns and can capture micro-level spatial heterogeneity.

[0033] Specifically, in one implementation of this embodiment, step S100 includes the following steps: Step S101: Obtain shared bicycle order data within the city limits; Step S102: Obtain the travel flow trajectory based on the shared bicycle order data; Step S103: Construct the cycling flow network based on the travel flow trajectory; Step S104: Based on the cycling flow network, evaluate the cycling flow similarity using the flow similarity measurement method, and identify the spatial distribution pattern using the hierarchical clustering algorithm.

[0034] In one implementation of this embodiment, step S104 includes the following steps: Step S104a: For any two cycling flow paths in the cycling flow network, construct a dynamically adjustable spatial buffer zone centered on the starting point and ending point of each cycling flow path. Step S104b: When the starting point and ending point of the two cycling flow paths fall into the spatial buffer at the same time, the two cycling flow paths are determined to be spatially similar flow paths.

[0035] In this embodiment, during the process of identifying spatial distribution patterns (i.e., cycling flow patterns), it is first necessary to perform flow similarity measurement, specifically as follows: First, the shared bicycle trip order data is converted into origin-destination (OD) flow objects. For example... Figure 3 As shown, Figure 3 Figure (a) shows the spatial distribution visualization of shared bicycle traffic in a certain city. To identify the similarity between cycling flows, this embodiment constructs a dynamically adjusted spatial buffer zone centered on the starting and ending points of any two cycling flow paths. This spatial buffer zone is used to determine the spatial proximity of the two flows. The radius of this spatial buffer zone is dynamically set according to the length of the flow path to avoid pattern recognition bias caused by differences in path scale. Specifically, when the starting and ending points of two cycling flows both fall within each other's buffer zones, they are considered to be spatially similar flows.

[0036] Figure 3 Figure (b) illustrates the spatial similarity relationships between different flows. Let any cycling flow be an example. Its starting point and ending point are respectively Construct radii of at the starting and ending points respectively. A circular buffer zone is defined. If the starting and ending points of two cycling flows simultaneously fall into this buffer zone, then the two flows are considered similar. The formula for determining the spatial similarity between two flows is: ; ; ; In the formula, It is a Boolean value, when and If the condition is true, it means that the two flows are similar. This represents the Euclidean distance between two points; The flow buffer radius; The length of each flow path; As an adjustment factor, it is set to 0.3 in this embodiment; In this embodiment, the maximum buffer radius is set to 500 meters, referencing the typical service radius of shared bicycles. This dynamic buffer radius mechanism can balance the recognition accuracy and generalization ability of flows at different scales, effectively supporting subsequent flow clustering and spatial pattern recognition.

[0037] In this embodiment, after performing flow similarity measurement, the spatial distribution pattern is identified using the hierarchical clustering algorithm; wherein, the spatial distribution pattern includes: clustering pattern, diverging pattern, and converging pattern; specifically, step S104 further includes the following steps: Step S104c: Initialize each of the cycling flow paths as an independent flow cluster; Step S104d: Calculate the midpoint position of each cycling flow path, construct a KD tree spatial index structure based on all midpoint positions, and merge flow clusters from bottom to top based on the KD tree spatial index structure to identify the clustering pattern. Step S104e: Select a set of cycling flow paths with adjacent starting points and random directions based on the starting point proximity discrimination method, and identify the divergence pattern; Step S104f: Select a set of cycling flow paths with adjacent endpoints and similar directions based on the endpoint proximity discrimination method, and identify the convergence pattern.

[0038] In this embodiment, to improve the efficiency and accuracy of cycling flow clustering, a hierarchical clustering algorithm based on KD tree optimization is adopted to achieve spatial clustering pattern recognition of cycling flow.

[0039] The method for identifying aggregation patterns is as follows: First, each cycling flow is initialized as an independent flow cluster, serving as the basic unit for bottom-up aggregation. Second, the midpoint position of each flow is calculated, and a KD-tree spatial index structure is constructed based on these midpoints to support efficient neighborhood search. Then, for each flow, a cluster is formed with its midpoint as the center and a radius of... Perform a KD-tree range query to obtain the set of adjacent flow combinations; based on this, traverse all flow pairs, and if they belong to different clusters, determine whether they meet the similarity discrimination criteria. If the conditions are met, the corresponding clusters will be merged into a new cluster.

[0040] After the above process is completed, the flow clusters are gradually merged from bottom to top, ultimately generating spatially close, uniformly oriented clusters of flow. The KD-tree optimized hierarchical clustering method used in this embodiment not only maintains high computational efficiency and stability under large-scale cycling flow data, but also significantly improves the clustering results' ability to accurately identify spatial patterns of flow, providing reliable basic data for subsequent urban cycling flow feature analysis and built environment correlation research.

[0041] The methods for identifying divergence and convergence patterns are as follows: This embodiment further identifies divergent and converging flow patterns. A divergent pattern refers to a group of flows with adjacent starting points and random directions; a converging pattern refers to a group of flows with adjacent ending points and converging directions. The identification strategy for both is as follows: ; In the formula, The maximum buffer radius is set to 250 meters in this embodiment to limit the cluster size. (Boolean value) and By determining whether the starting or ending points of cycling flows converge, this method identifies flow clusters exhibiting either divergent or convergent characteristics. It eliminates the need for directional clustering, making it suitable for pattern extraction under irregular directional changes and effectively compensating for the shortcomings of traditional pattern recognition methods in adapting to heterogeneous flows.

[0042] Table 1. Identification methods and characteristics of three spatial flow patterns:

[0043] This embodiment proposes an identification strategy based on flow similarity measurement and KD-tree optimized hierarchical clustering, which can effectively extract typical cycling flow patterns such as clustering, diverging, and converging. Compared with existing studies that are mostly limited to the identification of single or limited patterns, this method realizes a systematic analysis of multiple types of travel patterns, thereby comprehensively characterizing the micro-level travel behavior features of cities and laying a technical foundation for multi-dimensional traffic pattern research.

[0044] like Figure 1 As shown, this embodiment of the invention provides a method for dynamically extrapolating urban population activity and spatial structure based on single-vehicle flow, including the following steps: Step S200: Aggregate the identified spatial distribution patterns according to the number of grid cells to construct a spatial level distribution model.

[0045] In this embodiment, spatial distribution patterns such as clustering, divergence, and convergence are identified, and the identified cycling flow patterns are aggregated according to grid units to construct a spatial hierarchy distribution model. This model can intuitively reflect the intensity and type differences of cycling characteristics in different regions, providing a reliable spatial basis for subsequent urban built environment analysis.

[0046] Specifically, in one implementation of this embodiment, step S200 includes the following steps: Step S201: Using a gridded spatial statistical method, the total number of starting points and ending points of cycling flows within the grid under the aggregation mode is counted, the number of starting points under the divergence mode is counted, and the number of ending points under the convergence mode is counted to obtain the scale level of each mode. Step S202: Construct the spatial level distribution model based on the statistical scale levels.

[0047] In this embodiment, after obtaining three cycling flow patterns (clustering, diverging, and converging), a gridded spatial statistical method is used to quantitatively analyze the spatial clustering characteristics of each pattern. Specifically, a regular grid of 250m × 250m is used to cover the study area, and differentiated statistics are performed for different patterns: the clustering pattern counts the total number of starting and ending points of the cycling flow within the grid, the diverging pattern only counts the number of starting points, and the converging pattern only counts the number of ending points.

[0048] After obtaining the statistical results, this embodiment uses the natural breakpoint classification method for grading, as shown in Table 2.

[0049] Table 2. Classification of Bicycle Sharing Traffic Intensity Levels:

[0050] Compared to equidistant or quantile-based grading methods, this method maximizes intra-class homogeneity while enhancing inter-class diversity, more realistically reflecting the spatial distribution characteristics of cycling flow. Combining the overall numerical distribution of the three patterns, the grid is divided into scale levels 0–5, and a spatial grading map is generated accordingly. Level 0 indicates no significant pattern; levels 1–2 correspond to low to medium activity areas; level 3 represents sub-centers or secondary travel hotspots; and levels 4–5 point to high-activity areas such as transportation hubs, large commercial areas, or high-density residential areas.

[0051] This method not only reveals the differences in aggregation intensity of different cycling flow patterns across modes, but also allows for further coupling with built environment data, providing scientific support for analyzing pattern formation mechanisms and revealing the evolutionary patterns of urban functional zones.

[0052] In this embodiment, a regular grid (250m × 250m) is introduced as the basic spatial unit. The starting point, ending point, or a combination of both are statistically analyzed for different patterns. Combined with the natural breakpoint classification method, the aggregation results are divided into 0–5 scale levels, and a spatial distribution map is generated (e.g., ...). Figure 4 (As shown). Compared with traditional studies that rely on administrative regions or traffic zones as spatial units, this method can reveal travel differences at a finer scale, achieve cross-modal dimension clustering intensity comparison, and significantly improve the precision and explanatory power of cycling flow spatial analysis.

[0053] like Figure 1 As shown, this embodiment of the invention provides a method for dynamically extrapolating urban population activity and spatial structure based on single-vehicle flow, including the following steps: Step S300: Combine the urban built environment characteristics with the spatial level distribution in the spatial level distribution model to construct a comprehensive dataset suitable for machine learning regression modeling.

[0054] In this embodiment, after constructing the spatial hierarchy distribution model, the characteristics of the urban built environment are combined with the spatial hierarchy distribution to construct a comprehensive dataset suitable for machine learning regression modeling.

[0055] Specifically, in one implementation of this embodiment, step S300 includes the following steps: Step S301: Construct a dataset containing urban built environment characteristics and associate it with the spatial levels in the spatial level distribution model to obtain the comprehensive dataset; wherein, the urban built environment characteristics include: urban density characteristics, urban design characteristics, urban diversity characteristics, public transportation accessibility characteristics, and destination accessibility characteristics.

[0056] In this embodiment, after completing the spatial aggregation analysis of cycling flow patterns, a pattern causal analysis method based on interpretable machine learning is further proposed. This method first constructs a dataset containing multi-dimensional built environment features and then models its association with the hierarchical zoning results of cycling flow patterns. The built environment feature system covers five dimensions: density (building density, floor area ratio), design (road density, intersection density), diversity (land use mixing), public transport accessibility (distance to subway station, distance to bus stop), and destination accessibility (transportation POI accessibility, work POI accessibility, leisure POI accessibility, and residential POI accessibility; POI refers to the smallest data unit in a geographic information system, containing information such as name, coordinates, and address), comprehensively reflecting urban spatial structure elements.

[0057] like Figure 1 As shown, this embodiment of the invention provides a method for dynamically extrapolating urban population activity and spatial structure based on single-vehicle flow, including the following steps: Step S400: Based on the comprehensive dataset, and combined with the machine learning model interpretation method of game theory, quantitatively analyze the nonlinear influence mechanism of the urban built environment characteristics on different spatial distribution patterns, and deduce the intrinsic correlation between urban population activities and spatial structure.

[0058] In this embodiment, based on the constructed comprehensive dataset and the introduction of an interpretable machine learning model, combined with the SHAP method, the nonlinear influence mechanism of urban built environment variables on different cycling flow patterns is quantitatively analyzed, revealing the intrinsic correlation between complex spatial structure and environmental characteristics, thereby enhancing the explanatory power of the causes of flow pattern generation and the interpretability of the model.

[0059] Specifically, in one implementation of this embodiment, step S400 includes the following steps: Step S401: Based on the comprehensive dataset, a regression model is established using a machine learning algorithm with nonlinear fitting capabilities; Step S402: Based on the regression model, use the machine learning model interpretation method of game theory to quantitatively analyze the contribution of the urban built environment characteristics to the prediction of different spatial distribution patterns. Step S403: Based on the predicted contribution, output the global feature importance ranking and local feature dependency map of the urban built environment features to obtain the inference result of the intrinsic correlation between urban population activities and spatial structure.

[0060] In this embodiment, to capture complex nonlinear relationships that are difficult to characterize using traditional statistical methods, machine learning algorithms with nonlinear fitting capabilities are employed to establish regression models. These include models such as Random Forest, Gradient Boosting Tree (GBDT), and XGBoost (an ensemble learning algorithm based on gradient boosting decision trees). In practical applications, the XGBoost model demonstrates superior predictive performance and can effectively support the interpretation and analysis of cycling flow patterns.

[0061] Regarding interpretability, this invention innovatively introduces the SHAP interpretation framework, which quantifies the contribution of each built environment feature to the model's predictions, thereby achieving a visual deconstruction of the prediction results. This method utilizes cooperative game theory to decompose the overall prediction into the marginal contributions of each feature, thus revealing the nonlinear influence mechanism between the built environment and different cycling flow patterns. The formula is expressed as follows: ; In the formula, The baseline output value is the model's predicted value when there are no feature inputs. Representation of features Contribution to the model. It is the number of features that the model actually inputs. It is an indicator variable, when the characteristic A value of 1 is assigned when included in the calculation, and 0 otherwise. When calculating the complete predictions of the model, all features... All values ​​are 1.

[0062] To improve computational efficiency, this embodiment further employs the TreeSHAP optimization algorithm (TreeSHAP is a Shapley value calculation algorithm specifically optimized for decision tree models; Shapley value is a mathematical method for fairly distributing the results of collaboration). Leveraging the structural characteristics of the tree model, the computational complexity is reduced from exponential to polynomial time, ensuring the feasibility of large-scale spatial data analysis. Based on the above methods, this embodiment ultimately outputs two types of interpretation results: (1) Ranking the importance of global features to clarify the overall influence of each built environment element under different flow patterns; (2) Local feature dependency map reveals the nonlinearity and threshold effect between a single environmental element and the flow pattern level.

[0063] The technical solution provided in this embodiment breaks through the limitations of traditional statistical models in explaining complex spatial travel flow patterns, and at the same time provides a quantitative decision-making basis for urban transportation planning. It can be used to identify key environmental factors that promote specific cycling flow patterns, so as to achieve scientific planning and refined management.

[0064] This embodiment constructs an analytical framework for urban dynamic structure based on the perspective of "flow space." By analyzing the hierarchical aggregation of different flow patterns and their correlation with built environment elements, it can not only reveal the functional differentiation of urban spatial structure but also depict its dynamic evolution patterns at different times. Compared with traditional studies that rely on static spatial morphology, this method breaks through by demonstrating the transformation process of urban functional areas in the time dimension, providing new theoretical tools and technical paths for urban planning and refined management.

[0065] Overall, this embodiment demonstrates through empirical application of shared bicycle travel data that the proposed method can not only automatically identify three typical travel modes—clustered, divergent, and convergent—but also systematically reveal their quantitative evolution and spatial distribution patterns at different times, accurately reflecting the differentiated characteristics of peak and off-peak periods. Spatially, it can characterize the cluster size and travel radius differences of different modes, identifying the coexistence of short-distance localized travel and long-distance connectivity. Temporally, it can reveal significant differences in travel intensity and spatial patterns between weekdays and weekends. Regarding causal explanation, by combining machine learning and the SHAP method, it can quantitatively identify the differentiated influence mechanisms of factors such as accessibility of transportation facilities, residential and employment distribution, and road network structure on different travel modes, effectively overcoming the shortcomings of traditional methods in complex nonlinear interpretations.

[0066] In summary, the method provided in this embodiment not only realizes the intelligent identification and interpretable analysis of shared bicycle travel patterns, but also provides scientific support for urban traffic planning and refined management, and has outstanding application value and promotion potential.

[0067] This embodiment achieves the following technical effects through the above technical solution: 1) This embodiment systematically models and spatially analyzes the shared bicycle travel flow. Based on the spatial distribution characteristics and the interaction between the starting points, the flow patterns of shared bicycles can be divided into clustered, divergent, and convergent types. An interpretable machine learning model is introduced to quantify the nonlinear influence of urban built environment characteristics on different flow patterns.

[0068] 2) This embodiment can not only reveal the urban polycentric structure and functional distribution characteristics reflected by different types of flow patterns, but also realize the dynamic deduction of urban spatial structure based on the spatiotemporal evolution of single-vehicle travel flow, depict the role transformation and spatial connection changes of the region at different times, thereby showing the evolution law of urban spatial pattern.

[0069] 3) This embodiment can overcome the shortcomings of existing technologies in pattern recognition and formation mechanism explanation, and provide a basis for predicting the evolution trend of urban functional areas, optimizing the layout of the built environment, promoting sustainable travel, and promoting balanced urban spatial development.

[0070] Exemplary device Based on the above embodiments, the present invention also provides a dynamic simulation system for urban crowd activity and spatial structure based on single-vehicle flow, comprising: The spatial distribution pattern recognition module is used to construct a cycling flow network. Based on shared bicycle order data within the city, it identifies spatial distribution patterns by combining flow similarity measurement methods with hierarchical clustering algorithms. The spatial hierarchy distribution model construction module is used to aggregate the identified spatial distribution patterns according to the number of grid cells to construct a spatial hierarchy distribution model. The comprehensive dataset construction module is used to combine the characteristics of the urban built environment with the spatial hierarchy distribution in the spatial hierarchy distribution model to construct a comprehensive dataset suitable for machine learning regression modeling. The cycling flow pattern interpretation module is used to quantitatively analyze the nonlinear influence mechanism of the urban built environment characteristics on different spatial distribution patterns based on the comprehensive dataset and combined with the machine learning model interpretation method of game theory, and to deduce the intrinsic correlation between urban population activities and spatial structure.

[0071] This embodiment achieves the following technical effects through the above technical solution: 1) This embodiment systematically models and spatially analyzes the shared bicycle travel flow. Based on the spatial distribution characteristics and the interaction between the starting points, the flow patterns of shared bicycles can be divided into clustered, divergent, and convergent types. An interpretable machine learning model is introduced to quantify the nonlinear influence of urban built environment characteristics on different flow patterns.

[0072] 2) This embodiment can not only reveal the urban polycentric structure and functional distribution characteristics reflected by different types of flow patterns, but also realize the dynamic deduction of urban spatial structure based on the spatiotemporal evolution of single-vehicle travel flow, depict the role transformation and spatial connection changes of the region at different times, thereby showing the evolution law of urban spatial pattern.

[0073] 3) This embodiment can overcome the shortcomings of existing technologies in pattern recognition and formation mechanism explanation, and provide a basis for predicting the evolution trend of urban functional areas, optimizing the layout of the built environment, promoting sustainable travel, and promoting balanced urban spatial development.

[0074] Based on the above embodiments, the present invention also provides a terminal, the principle block diagram of which can be as follows: Figure 5 As shown.

[0075] The terminal includes: a processor, a memory, an interface, a display screen, and a communication module connected via a system bus; wherein, the processor of the terminal provides computing and control capabilities; the memory of the terminal includes a computer-readable storage medium and internal memory; the computer-readable storage medium stores an operating system and computer programs; the internal memory provides an environment for the operation of the operating system and computer programs in the computer-readable storage medium; the interface is used to connect to external devices; the display screen is used to display relevant information; and the communication module is used to communicate with a cloud server or other devices.

[0076] When executed by the processor, this computer program is used to implement a method for dynamically extrapolating urban population activity and spatial structure based on single-vehicle flow.

[0077] It will be understood by those skilled in the art that Figure 5 The schematic diagram shown is merely a partial structural diagram related to the present invention and does not constitute a limitation on the terminal to which the present invention is applied. A specific terminal may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0078] In one embodiment, a terminal is provided, comprising: a processor and a memory, the memory storing a program for dynamically extrapolating urban population activity and spatial structure based on single-vehicle flow, the program being executed by the processor to implement the operation of the above-described method for dynamically extrapolating urban population activity and spatial structure based on single-vehicle flow.

[0079] In one embodiment, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a dynamic simulation program for urban crowd activity and spatial structure based on single-vehicle flow, which, when executed by a processor, is used to implement the operation of the above-described dynamic simulation method for urban crowd activity and spatial structure based on single-vehicle flow.

[0080] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, database, or other media used in the embodiments provided by this invention can include both non-volatile and volatile memory.

[0081] In summary, this invention provides a method, system, terminal, and storage medium for dynamically extrapolating urban population activities and spatial structure based on bicycle flow. The method includes: constructing a cycling flow network; identifying spatial distribution patterns based on shared bicycle order data within the city using a flow similarity measurement method combined with a hierarchical clustering algorithm; aggregating the identified spatial distribution patterns according to grid units to construct a spatial hierarchy distribution model; combining urban built environment characteristics with the spatial hierarchy distribution in the spatial hierarchy distribution model to construct a comprehensive dataset suitable for machine learning regression modeling; and based on the comprehensive dataset, using game theory-based machine learning model interpretation methods to quantitatively analyze the nonlinear influence mechanism of urban built environment characteristics on different spatial distribution patterns, and extrapolate the intrinsic correlation between urban population activities and spatial structure. This invention comprehensively realizes the dynamic evolution process of urban spatial structure.

[0082] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A method for dynamically extrapolating urban population activity and spatial structure based on single-vehicle flow, characterized in that, include: A cycling flow network is constructed based on shared bicycle order data within the city, and spatial distribution patterns are identified by combining flow similarity measurement methods with hierarchical clustering algorithms. The identified spatial distribution patterns are aggregated according to the number of grid cells to construct a spatial hierarchical distribution model; By combining the characteristics of the urban built environment with the spatial hierarchy distribution in the aforementioned spatial hierarchy distribution model, a comprehensive dataset suitable for machine learning regression modeling is constructed. Based on the comprehensive dataset, and combined with the machine learning model interpretation method of game theory, the nonlinear influence mechanism of the urban built environment characteristics on different spatial distribution patterns is quantitatively analyzed, and the intrinsic correlation between urban population activities and spatial structure is deduced.

2. The method for dynamic extrapolation of urban population activity and spatial structure based on single-vehicle flow according to claim 1, characterized in that, The construction of the cycling flow network, based on shared bicycle order data within the city, identifies spatial distribution patterns using a flow similarity measurement method combined with a hierarchical clustering algorithm, including: Obtain shared bicycle order data within the city limits; The travel flow trajectory is obtained based on the shared bicycle order data; The cycling flow network is constructed based on the travel flow trajectory. Based on the cycling flow network, the similarity of cycling flows is evaluated using the flow similarity measurement method, and the spatial distribution pattern is identified using the hierarchical clustering algorithm.

3. The method for dynamic extrapolation of urban population activity and spatial structure based on single-vehicle flow according to claim 2, characterized in that, The method of evaluating cycling flow similarity using the flow similarity metric includes: For any two cycling paths in the cycling flow network, a dynamically adjustable spatial buffer zone is constructed with the starting point and ending point of each cycling flow path as the center. When the starting point and ending point of two cycling paths simultaneously fall into the spatial buffer zone, the two cycling paths are determined to be spatially similar flow paths.

4. The method for dynamic extrapolation of urban population activity and spatial structure based on single-vehicle flow according to claim 2, characterized in that, The spatial distribution patterns include: clustering pattern, divergence pattern, and convergence pattern; The method of identifying the spatial distribution pattern using the hierarchical clustering algorithm includes: Initialize each cycling flow path as an independent flow cluster; Calculate the midpoint position of each cycling flow path, construct a KD tree spatial index structure based on all midpoint positions, and merge flow clusters from bottom to top based on the KD tree spatial index structure to identify the clustering pattern; Based on the starting point proximity discrimination method, a set of cycling flow paths with adjacent starting points and random directions is selected, and the divergence pattern is identified. Based on the endpoint proximity discrimination method, a set of cycling flow paths with adjacent endpoints and similar directions is selected, and the convergence pattern is identified.

5. The method for dynamic extrapolation of urban population activity and spatial structure based on single-vehicle flow according to claim 1, characterized in that, The process of aggregating the identified spatial distribution patterns according to the number of grid cells to construct a spatial hierarchy distribution model includes: Using a gridded spatial statistical method, the total number of starting points and ending points of cycling flows within the grid under the clustering mode is counted, the number of starting points under the diverging mode is counted, and the number of ending points under the converging mode is counted to obtain the scale level of each mode. The spatial hierarchy distribution model is constructed based on the statistically analyzed scale levels.

6. The method for dynamic extrapolation of urban population activity and spatial structure based on single-vehicle flow according to claim 1, characterized in that, The method of combining urban built environment characteristics with the spatial hierarchy distribution in the spatial hierarchy distribution model to construct a comprehensive dataset suitable for machine learning regression modeling includes: A dataset containing urban built environment characteristics is constructed and correlated with the spatial levels in the spatial level distribution model to obtain the comprehensive dataset; wherein, the urban built environment characteristics include: urban density characteristics, urban design characteristics, urban diversity characteristics, public transportation accessibility characteristics, and destination accessibility characteristics.

7. The method for dynamic extrapolation of urban population activity and spatial structure based on single-vehicle flow according to claim 1, characterized in that, Based on the comprehensive dataset, and using a game theory-based machine learning model interpretation method, the nonlinear influence mechanism of the urban built environment characteristics on different spatial distribution patterns is quantitatively analyzed, and the intrinsic correlation between urban population activities and spatial structure is deduced, including: Based on the comprehensive dataset, a regression model is established using a machine learning algorithm with nonlinear fitting capabilities; Based on the regression model, the contribution of the urban built environment characteristics to the prediction of different spatial distribution patterns is quantitatively analyzed using a game theory-based machine learning model interpretation method. Based on the predicted contribution, the global feature importance ranking and local feature dependency map of the urban built environment features are output, and the inference results of the intrinsic correlation between urban population activities and spatial structure are obtained.

8. A dynamic simulation system for urban crowd activity and spatial structure based on single-vehicle flow, characterized in that, include: The spatial distribution pattern recognition module is used to construct a cycling flow network. Based on shared bicycle order data within the city, it identifies spatial distribution patterns by combining flow similarity measurement methods with hierarchical clustering algorithms. The spatial hierarchy distribution model construction module is used to aggregate the identified spatial distribution patterns according to the number of grid cells to construct a spatial hierarchy distribution model. The comprehensive dataset construction module is used to combine the characteristics of the urban built environment with the spatial hierarchy distribution in the spatial hierarchy distribution model to construct a comprehensive dataset suitable for machine learning regression modeling. The cycling flow pattern interpretation module is used to quantitatively analyze the nonlinear influence mechanism of the urban built environment characteristics on different spatial distribution patterns based on the comprehensive dataset and combined with the machine learning model interpretation method of game theory, and to deduce the intrinsic correlation between urban population activities and spatial structure.

9. A terminal, characterized in that, include: The processor and memory, wherein the memory stores a dynamic simulation program for urban population activity and spatial structure based on single-vehicle flow, and the dynamic simulation program for urban population activity and spatial structure based on single-vehicle flow is executed by the processor to implement the operation of the dynamic simulation method for urban population activity and spatial structure based on single-vehicle flow as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a dynamic simulation program for urban population activity and spatial structure based on single-vehicle flow. When executed by a processor, the dynamic simulation program for urban population activity and spatial structure based on single-vehicle flow is used to implement the operation of the dynamic simulation method for urban population activity and spatial structure based on single-vehicle flow as described in any one of claims 1-7.

Images