Life circle identification and full-age adaptive facility supply and demand evaluation method based on resident flow

Abstract

The invention relates to the technical field of urban public service facility evaluation and space planning, and discloses a living circle identification and full-age adaptive facility supply and demand evaluation method based on resident flow. The core of the method is to dynamically capture an actual travel track of a resident by using mobile phone signaling data so as to identify a life circle range highly matched with an actual behavior; a two-dimensional demand weight matrix of groups and facilities according to age groups is constructed, and differentiated demands of different age groups such as children, young adults and old people for various public service facilities are finely quantified; by fusing multi-source data such as population distribution, facility point locations and real-time road networks, accurate matching evaluation of facility supply and demand services is realized by adopting an improved space interaction model; according to the method, the limitation that traditional planning depends on a static administrative boundary is effectively overcome, life circle division is more scientific, and the fineness and fairness of facility configuration are remarkably improved from the perspective of all-age demand evaluation.

Description

Technical Field

[0001] This invention relates to the field of urban public service facility assessment and spatial planning technology, specifically to a method for identifying residents' living circles based on their mobility and assessing the supply and demand of facilities suitable for all ages. Background Technology

[0002] With the accelerating pace of urbanization, urban residents' demands for public service facilities are becoming increasingly diversified and sophisticated. Traditional urban planning methods rely heavily on static population data and fixed administrative boundaries for facility layout, making it difficult to dynamically capture residents' actual activity patterns and spatial behavior characteristics. This results in a significant mismatch between facility supply and residents' actual needs, and the problem of both facility redundancy and shortage is becoming increasingly prominent in some areas.

[0003] In recent years, advancements in Geographic Information Systems (GIS) and location-based services (Location-Based Services) have provided crucial tools for the spatial analysis of public service facilities. Researchers have successively proposed methods such as the shortest path method, the two-step movement search method, and gravity models to assess the spatial accessibility of facilities. However, these methods are still largely based on static assumptions and fail to fully incorporate residents' actual travel trajectories and dynamic activity patterns. Meanwhile, with the widespread adoption of mobile communication technology, mobile phone signaling data provides a new data foundation for characterizing population flow patterns. Existing studies have attempted to utilize this data to optimize facility layout, but significant limitations remain: on the one hand, existing methods for dividing living areas largely rely on fixed spatiotemporal radii, failing to accurately reflect the complexity and diversity of residents' activities; on the other hand, most studies lack a systematic consideration of the differentiated needs of people across all age groups, making it difficult to provide precise facility configuration recommendations for different groups such as children, young adults, and the elderly.

[0004] Therefore, there is an urgent need for a public service facility evaluation method that can deeply integrate residents' dynamic travel data, scientifically identify the scope of living circles, and systematically embed the needs characteristics of all ages, so as to improve the scientific nature, fairness, and suitability of urban facility allocation. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method for identifying residential circles based on resident mobility and assessing the supply and demand of facilities suitable for all ages. This method has the advantages of high accuracy and efficiency, and solves the aforementioned technical problems.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for identifying residential circles based on resident mobility and assessing the supply and demand of facilities suitable for all ages, comprising the following steps: S1: Collect mobile phone signaling data, street data, road network and water system data, population distribution data by age group, public service facility location data and relevant planning standards in the study area and preprocess them respectively; S2: Extract resident stop points from the preprocessed mobile signaling data, associate resident stop points with the corresponding blocks, and extract the flow intensity between stop points within the blocks; S3: Perform spatial modeling of street data and flow intensity to construct a spatial topology map and pedestrian flow relationship map of the spatial modeling data; S4: Embedding learning is performed on the people flow relationship graph. The embedded people flow relationship graph and the spatial topology graph are then used to perform graph clustering using the CESNA algorithm. After multiple iterations, the living circle affiliation label is obtained. S5: Calculate the differentiated service distance for each living circle based on the preprocessed population distribution data by age group; S6: Construct the demand weight matrix of the preprocessed population distribution data by age group, and obtain the demand level of the population in all age groups based on the demand weight matrix. S7: Based on the preprocessed road network and water system data, public service facility location data, relevant planning standards, and preprocessed age-group population distribution data and service distance of the living circle, generate the service range of public service facilities; S8: Based on the service scope of public service facilities and the demand level of the population of all ages, the overall supply and demand matching efficiency is obtained.

[0007] As a preferred technical solution of the present invention, the S1 street data is obtained by identifying closed spatial units formed by the enclosed spatial elements based on urban road networks and water systems with physical boundary functions in a GIS platform, and treating the closed spatial unit data as street data.

[0008] As a preferred embodiment of the present invention, the S3 spatial modeling includes the following steps: S3.1: Read the latitude and longitude coordinates of the geometric center point of the street block area layer generated in ArcGIS Pro; S3.2: Based on the preprocessed latitude and longitude coordinates of the block center points, the great circle distance between any two block center points is calculated using the Haversine formula. The specific calculation formula is as follows: In the formula, R represents the distance between the center points of two blocks; R is the Earth's radius. , These are the latitudes of the two points; , These are the longitudes of the two points; It is the arcsine function; It is a sine function; It is a cosine function; S3.3: Model using the geopy, networkx, and scikit-learn libraries to obtain the distance matrix.

[0009] As a preferred technical solution of the present invention, S3 constructs a spatial topology map and a human flow relationship map of spatial modeling data; The spatial topology graph is constructed using the NearestNeighbors algorithm from the scikit-learn library based on the distance matrix. The pedestrian flow graph uses street blocks as nodes. Based on pedestrian flow data between blocks, the edge weights are calculated using the weighted average of weekday and weekend flow intensities to construct the topological structure of the pedestrian flow graph, which is stored as an adjacency matrix. ; The formula for calculating flow strength is: In the formula, This indicates the flow intensity from block a to block b; This indicates the frequency of movement from block a to block b; This represents the average distance traveled from block a to block b.

[0010] As a preferred technical solution of the present invention, step S4 performs embedding learning on the people flow relationship graph, and calls the CESNA algorithm to perform graph clustering on the embedded people flow relationship graph and the spatial topology graph, obtaining the living circle affiliation label after multiple iterations, including the following steps: S4.1: Embedding learning is performed on the pedestrian flow graph using a graph convolutional network. During the propagation process of the graph convolutional network, the following formula is used for the hierarchical transfer of node features: In the formula, To represent the first in the network The node feature matrix of the layer; Indicates the first in the network The node feature matrix of the layer; For activation functions; Adjacency matrix The enhanced adjacency matrix formed after adding self-connections; yes The degree matrix; Indicates the first The trainable weight matrix of the layer; S4.2: Use the CESNA algorithm to perform graph clustering on the embedded pedestrian flow graph and spatial topology graph, randomly initialize the community membership matrix, and update the community membership matrix in each iteration. The community membership update formula of the CESNA algorithm is as follows: In the formula, Indicates the first One block; Indicates the first One community; Indicates a street Belongs to the community Membership degree; The membership component based on topology is obtained by calculating the structural similarity of blocks in the spatial topology graph using the networkx library. The membership component based on neighborhood attributes is obtained by calculating the cosine similarity between the neighborhood embedding vector and the community centroid of the embedded pedestrian flow graph using the NumPy library. S4.3: After a fixed number of iterations and convergence, determine the final living circle affiliation label of each block based on its maximum community affiliation degree; In the formula, Indicates the first One block; Indicates the first A living circle; For the first The final living circle affiliation label of each neighborhood; To represent taking the option that maximizes the function value value; After a fixed number of iterations, the th... The block belongs to the first Community affiliation of each living circle.

[0011] As a preferred embodiment of the present invention, the formula for calculating the service distance between the preprocessed age-group population distribution data and the living circle in step S5 is as follows: In the formula, For the living circle Suitable walking distance; , , , respectively living circle The proportions of young people, middle-aged people, teenagers and the elderly.

[0012] As a preferred embodiment of the present invention, step S6 constructs a demand weight matrix of the preprocessed population distribution data by age group, and obtains the demand level of the entire population by age group based on the demand weight matrix, including the following steps: S6.1: The formula for the demand weight matrix of the preprocessed age-group population distribution data is as follows: In the formula, express Age groups' demand for public service facilities Demand weight, express Public service facilities for different age groups The total number of people in the grid; Indicates the study area Total population of the same age group; S6.2: Construct a two-dimensional demand weight matrix for age groups and facilities based on the demand weight matrix formula; S6.3: Obtain the population demand level for all age groups based on the demand weight matrix, using the following formula: in, Indicates living circle Internal public service facilities The total age-weighted demand level of the population; Indicates living circle Inside Population size of the relevant age group; express Age groups' demand for public service facilities Demand weight; To sum the information from the four age groups.

[0013] As a preferred technical solution of the present invention, step S7 generates the service range of public service facilities based on preprocessed road network and water system data, public service facility data, relevant planning standards, and preprocessed age-group population distribution data and service distance of living circles, including the following steps: S7.1: In ArcGIS Pro, use the Network Analyst module to construct a road network dataset, abstracting roads into a network model containing nodes and edges, as shown in the following formula: In the formula, G represents the road network diagram; V represents a node of the road network, which consists of the coordinates of all road endpoints; and The horizontal and vertical coordinates represent the specific road endpoints; Represents a specific road endpoint; R represents the x-th road segment; R represents the set of the complete road network, which includes all road segments within the study area; E represents the set of edges of the road network, which consists of all continuous road segments. Representing an edge The weights are the Euclidean distances between the two endpoints, calculated by the L2 norm of the coordinate difference. get; To sum the road segments within the study area; Used as a separator; It is a norm 2; S7.2: Locate the facility point to the nearest node in the road network. The calculation formula is as follows: In the formula, Indicates public service facilities The nearest node in the road network Represents road nodes and public service facilities The Euclidean distance between them; To select the node corresponding to the minimum function value; S7.3: Using the nearest facility point analysis tool, starting from the location of public service facilities, and with the service distance between the preprocessed age-group population distribution data and the living circle as the search radius, reachable nodes and reachable edges are extracted. The calculation formula is as follows: In the formula, j represents a public service facility; Indicates The graph centered on the target is a reachable subgraph. Let j be the set of reachable nodes for public service facility j; Let j be the set of reachable edges of public service facility j; Represents a node To the node The shortest path distance; For the living circle Suitable walking distance; S7.4: Perform Delaunay triangulation on the reachable node set and filter by the circumcircle radius threshold to generate the service boundary of public service facilities. The calculation formula is as follows: In the formula, This represents the radius of the circumcircle of the Delaunay triangle; This is a preset threshold parameter; This is the corresponding radius threshold.

[0014] As a preferred embodiment of the present invention, step S8, based on the service scope of public service facilities and the demand levels of the population across all age groups, obtains the comprehensive supply and demand matching efficiency, including the following steps: S8.1: Based on the service scope of public service facilities, use spatial connectivity and summary statistics tools in ArcGIS to calculate the service coverage area of ​​various facilities within each living circle. The quality coefficient of public service facility supply is obtained using the following formula: In the formula, For the living circle Internal public service facilities The supply quality coefficient; Indicates living circle Internal public service facilities The service area; Indicates living circle Internal public service facilities The total amount; Indicates living circle area; Indicates the social circle Sum all public service facilities j; S8.2: The supply level of public service facilities is obtained based on the quality coefficient of public service facility supply, using the following formula: In the formula, Indicates living circle Internal public service facilities Supply level; Indicates living circle Internal public service facilities Quantity; Indicates public service facilities The plan targets a thousand-person quota; For the living circle Internal public service facilities The supply quality coefficient; S8.3: Based on the supply level of public service facilities, a comprehensive supply-demand matching index is obtained, as shown in the following formula: In the formula, For the living circle Chinese public service facilities The supply and demand matching index; For the living circle Internal public service facilities Supply level; For the living circle Internal public service facilities The total age-weighted demand level of the population; S8.4: Based on the comprehensive supply and demand matching index, the comprehensive supply and demand matching efficiency is obtained, as shown in the following formula: In the formula, For the living circle The overall spatial supply and demand matching efficiency of all public service facilities; Indicates living circle Internal public service facilities The total amount; Life Circle Chinese public service facilities The supply and demand matching index; To express one's love for one's living circle All public service facilities are summed.

[0015] Compared with existing technologies, this invention provides a method for identifying living circles based on resident mobility and assessing the supply and demand of facilities suitable for all ages, which has the following beneficial effects: 1. This invention breaks through the limitations of traditional static population data by using dynamic behavioral data such as mobile phone signaling data. It can accurately capture the actual activity characteristics of residents across administrative boundaries, realize the transformation of the living circle range from a fixed radius to a dynamic boundary, and significantly improve the scientificity and accuracy of living circle identification.

[0016] 2. This invention effectively quantifies the differentiated needs of different age groups for various facilities by using age-group population distribution and a two-dimensional facility demand weight matrix. It solves the problem that traditional methods do not adequately consider the needs of all ages, and shifts facility assessment from a one-size-fits-all extensive model to a refined management based on groups and categories.

[0017] 3. This invention establishes a complete facility assessment technology system by integrating road network analysis, spatial clustering and supply-demand matching algorithms, realizing full-process automation from data collection and processing to result output, greatly improving work efficiency and providing reliable decision support tools for urban planning departments.

[0018] 4. Through innovative methods such as multi-source data fusion, all-age demand modeling, and dynamic evaluation, a scientific, accurate, and operable facility evaluation system has been formed, which has good application prospects and promotion value. Attached Figure Description

[0019] Figure 1 This is a technical framework diagram of the present invention; Figure 2 Population distribution map; Figure 3 Identify the living circles in the central urban area of ​​Kunming; Figure 4 The supply and demand matching index for public service facilities; Figure 5 To improve the overall efficiency of public service facility space matching; Figure 6 For dynamic identification of urban living circles; Figure 7 This is the interface for adjusting the requirement weights of this invention; Figure 8 This module showcases the supply and demand matching assessment of public service facilities in urban living areas. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Please see Figures 1-8 The method for identifying residents' living circles based on their mobility and assessing the supply and demand of facilities suitable for all ages includes the following steps: S1: Collect mobile phone signaling data, road network and water system data, street data, population distribution data by age group, public service facility location data and relevant planning standards in the study area and preprocess them respectively; Mobile signaling data includes users' spatiotemporal location information, obtained through base station positioning to track residents' activity trajectories; road network data includes urban road grades, lengths, and topological relationships, used to construct a spatial analysis network; water system data uses boundary data of Dianchi Lake in Kunming; public service facility location data covers the geographic coordinates of five categories of facilities: commercial services, elderly care services, basic education, cultural and sports entertainment, and medical and health facilities; public service facility location data and related planning standards are obtained based on national and local standards such as the "Urban Residential Area Planning and Design Standard" (GB50180-2018) and the "Urban Public Service Facility Planning Standard". Planning parameters such as service radius, per capita index, and configuration requirements for various facilities; population distribution data by age group is based on 500m×500m grid population data obtained by inverting mobile phone signaling data, and the population is counted according to four age groups: 0-18 years old (teenagers), 19-35 years old (young adults), 36-59 years old (middle-aged), and 60 years old and above (elderly), forming the spatial distribution characteristics of the population across all ages; street block data is obtained in the GIS platform based on spatial elements with physical boundaries of urban road network and water system, identifying closed spatial units enclosed by spatial elements, and treating closed spatial unit data as street block data; Preprocessing includes preprocessing mobile phone signaling data, road network and water system data, street data, public service facility location data, and population distribution data by age group; Mobile signaling data preprocessing was completed in a Python 3.8 environment, using libraries such as pandas, numpy, and scipy for deep cleaning and processing. Specifically, this included: using the interquartile range method to identify and remove abnormal location points, setting a speed threshold of 300 km / h to filter unreasonable movement trajectories; using the Kalman filter algorithm to smooth and reconstruct residents' activity trajectories and fill in missing data points; and using a density-based spatiotemporal clustering algorithm to extract dwell points, setting a time threshold of 30 minutes and a spatial threshold of 500 meters to identify residents' stable dwelling locations and their duration. The preprocessing of road network and water system data was completed in the ArcGIS platform, including: performing topological checks on road network data, correcting hanging nodes and redundant line segments, and constructing a connected road network; performing boundary smoothing on the boundary data of Kunming Dianchi Lake, eliminating small jagged edges, and extracting the main water body boundary lines for street unit division. The preprocessing of public service facility location data includes: coordinate correction and standardization, unifying facility coordinates from different sources to the WGS84 coordinate system; attribute standardization, classifying and coding facilities into five categories: commercial services, elderly care services, basic education, cultural and sports entertainment, and medical and health care; and data integrity verification, supplementing missing attribute information. The preprocessing of public service facility location data and planning standard data mainly includes: digitizing paper-based specification documents and extracting key parameters such as per capita indicators and service radius for various facilities; and establishing a standard parameter database to provide a standardized basis for subsequent supply and demand matching analysis. In a Python 3.8 environment, the age-group population distribution data was preprocessed using pandas and numpy libraries. This included data cleaning and formatting, reading the raw grid population data, verifying the completeness of population data for the four age groups (0-18 years, 19-35 years, 36-59 years, and 60 years and above), handling missing values ​​using pandas, and ensuring that each 500m×500m grid cell contained complete age-group population statistics. This provided fundamental data support for subsequent living area identification and facility supply-demand matching analysis. All spatial data (street data, road network and water system data, public service facility location data, Kunming Dianchi Lake water boundary data, and 500m×500m grid population distribution areal data generated through mobile phone signaling data) were uniformly converted to the WGS84 coordinate system, and the data format was standardized to a unified vector data format. This standardization process ensures that spatial data from different data sources and coordinate benchmarks can be integrated and analyzed under a unified geographic reference framework, eliminating spatial location deviations caused by coordinate system differences. This provides an accurate and reliable data foundation for subsequent dual-map structure construction, network analysis, and spatial calculations, and is a technical prerequisite for realizing multi-source data collaborative analysis. S2: Extract resident stop points from the preprocessed mobile signaling data, associate resident stop points with the corresponding blocks, and extract the flow intensity between stop points within the blocks; The extraction of resident dwelling points uses a spatiotemporal clustering algorithm, with a time threshold of 30 minutes and a spatial threshold of 500 meters, to identify stable resident dwelling locations; spatial matching is used to associate resident dwelling points with their respective street blocks, employing point-area overlay analysis. Flow intensity is calculated based on the frequency and distance of residents' movements between different neighborhoods. The ratio of movement frequency to movement distance is used as a quantitative indicator of flow intensity. The calculation formula is as follows: In the formula, This indicates the flow intensity from block a to block b; This indicates the frequency of movement from block a to block b; This represents the average distance traveled from block a to block b. S3: Perform spatial modeling of street data and flow intensity to construct a spatial topology map and pedestrian flow relationship map of the spatial modeling data; S3 spatial modeling includes the following steps: S3.1: Read the latitude and longitude coordinates of the geometric center point of the street block area layer generated in ArcGIS Pro; S3.2: Based on the preprocessed latitude and longitude coordinates of the block center points, the Hasversine formula is used to calculate the great circle distance between any two block center points, generating an N×N dimensional distance matrix and storing it as a NumPy array. The specific calculation formula is as follows: In the formula, This represents the distance between the center points of two blocks, in km; R is the Earth's radius, taken as 6371 km. , These are the latitudes of the two points; , These are the longitudes of the two points; It is the arcsine function; It is a sine function; It is a cosine function; S3.3: Use libraries such as geopy, networkx, and scikit-learn to perform quantitative modeling of spatial relationships and obtain the distance matrix; The spatial topology graph is constructed using the NearestNeighbors algorithm from the scikit-learn library based on the distance matrix. The distance threshold is set to 1 kilometer, and each node can connect to a maximum of 10 neighboring nodes. The k-nearest neighbor algorithm is used to select neighboring nodes that meet the conditions. The constructed adjacency matrix is ​​symmetric, and the edge weights are set based on the normalized spatial distance. The pedestrian flow graph uses street blocks as nodes. The node feature matrix contains feature data of three dimensions: perimeter, area, and road network density of the street blocks. Edge weights are calculated based on the weighted average of weekday and weekend pedestrian flow intensity, using weekend and weekend pedestrian flow data between streets. The weight values ​​of all potential connecting edges are calculated, and then the 75th quantile is determined as a threshold. Connecting edges with weight values ​​below this threshold are discarded, and only connections with weight values ​​at or above the threshold are retained. Finally, the topology of the pedestrian flow graph is constructed based on the filtered connecting edges and stored as an adjacency matrix. This matrix records the significant connections between neighborhoods based on resident mobility, ensuring that the graph contains only significant resident mobility patterns. S4: Embedding learning is performed on the people flow relationship graph. The embedded people flow relationship graph and the spatial topology graph are then used to perform graph clustering using the CESNA algorithm. After multiple iterations, the living circle affiliation label is obtained. S4 performs embedding learning on the people flow relationship graph, and then uses the CESNA algorithm to perform graph clustering on the embedded people flow relationship graph and spatial topology graph. After multiple iterations, the community affiliation labels are obtained, including the following steps: S4.1: Implement a two-layer GCN network based on the PyTorch Geometric Graph Neural Network framework; the specific network configuration parameters are: first layer input dimension 3, output dimension 64; second layer input dimension 64, output dimension 32; using the ReLU activation function; employing the Adam optimizer with a learning rate set to 0.01; the number of training epochs is fixed at 200; the batch size is the entire graph data; during training, the adjacency matrix is ​​adjusted... Self-connects are added and symmetric normalization is performed to ensure the stability of feature propagation. During the propagation process of the graph convolutional network, the following formula is used for the hierarchical transfer of node features: In the formula, To represent the first in the network The node feature matrix of the layer; Indicates the first in the network The node feature matrix of a layer has a dimension determined by the number of nodes N (i.e., the number of blocks) and the feature dimension of that layer. The activation function is used to enhance the network's expressive power by introducing nonlinear transformations; Adjacency matrix ,Right now , where I is the identity matrix of the same order as A; yes The degree matrix is ​​a diagonal matrix; This represents the trainable weight matrix of the l-th layer, which is responsible for linearly transforming the features; S4.2: The CESNA algorithm is used to perform graph clustering on the embedded pedestrian flow graph and spatial topology graph. This algorithm is implemented using Python's networkx and numpy libraries, with the number of communities k=10, the maximum number of iterations T=100, and the community membership matrix and attribute-community weight matrix randomly initialized. In each iteration, the algorithm simultaneously considers the connectivity structure of the spatial topology graph and the node embedding features learned from GCN, updating these two matrices through an alternating optimization strategy. The community membership update formula for the CESNA algorithm is: In the formula, Indicates the first "A block" here specifically refers to a block. Indicates the first In this invention, "a community" specifically refers to a living circle identified through clustering. This indicates the degree to which block a belongs to community c; The membership component based on topology is obtained by calculating the structural similarity of nodes in the spatial topology graph using the networkx library. The membership component based on neighborhood attributes is obtained by calculating the cosine similarity between the neighborhood embedding vector and the community centroid of the embedded pedestrian flow graph using the NumPy library. S4.3: After a fixed number of iterations and convergence, determine the final living circle affiliation label of each block based on its maximum community affiliation degree; In the formula, Indicates the first One block; Indicates the first A living circle; For the first The final living circle affiliation label of each neighborhood; To represent taking the option that maximizes the function value value; After a fixed number of iterations, the th... The block belongs to the first Community affiliation of each living circle.

[0022] S5: Calculate the service distance between the preprocessed age-group population distribution data and the living circle; The age stratification uses four age groups: 0-18 years old (adolescents), 19-35 years old (young adults), 36-59 years old (middle-aged adults), and 60 years and above (elderly). The service radius is set based on walking speed and endurance time. For adolescents and the elderly, the service radius is set at 1000 meters based on a walking speed of 1.0 m / s and an endurance time of 15 minutes; for young adults and middle-aged adults, the service radius is set at 1500 meters based on a walking speed of 1.2 m / s and an endurance time of 20 minutes. Differential walking distances are calculated for each living circle i. Based on behavioral characteristic parameters of different age groups, and through population structure weighting, the differentiated service distances for each living circle are obtained. The calculation formula is as follows: In the formula, For the living circle Suitable walking distance; , , , respectively living circle The proportions of young people, middle-aged people, teenagers, and the elderly; S6: Construct the demand weight matrix of the preprocessed population distribution data by age group, and obtain the demand level of the population in all age groups based on the demand weight matrix. S6 constructs a demand weight matrix for the preprocessed age-group population distribution data, and obtains the demand level for all age groups based on the demand weight matrix, including the following steps: S6.1: In the ArcGIS Pro platform, spatial analysis tools were used to overlay and analyze 500m×500m grid population data generated from mobile phone signaling data with the base map of the study area. The proportion of each age group in each grid to the total population of that age group in the study area was calculated, and a two-dimensional demand weight matrix of age groups and facilities was constructed. This matrix quantifies the differentiated demand intensity of different age groups for five types of facilities: commercial services, elderly care services, basic education, cultural and sports entertainment, and medical and health care. This provides a data foundation for subsequent all-age demand assessment. The specific formula is as follows: In the formula, express Age groups' demand for public service facilities Demand weight, express Public service facilities for different age groups The total number of people in the grid; Indicates the study area S6.2: Construct a two-dimensional demand matrix for age groups and facilities based on the demand weight matrix formula; S6.3: Obtain the population demand level for all age groups based on the demand weight matrix, using the following formula: in, Indicates living circle Internal public service facilities The total age-weighted demand level of the population; Indicates living circle Inside Population size of the relevant age group; express Age groups' demand for public service facilities Demand weight; To sum the information from the four age groups; S7: Based on the preprocessed road network and water system data, public service facility location data, relevant planning standards, and preprocessed age-group population distribution data and service distance of the living circle, generate the service range of public service facilities; S7 generates the service area of ​​public service facilities based on preprocessed road network and water system data, public service facility location data, relevant planning standards, and preprocessed age-group population distribution data and service distances to living circles. This process includes the following steps: S7.1: In ArcGIS Pro, use the Network Analyst module to construct a road network dataset, abstracting roads into a network model containing nodes and edges, as shown in the following formula: In the formula, G represents the road network diagram; V represents the nodes of the road network, which consists of all road endpoints. This set is obtained by merging the endpoints of all road segments and is the basic node set for network analysis. and The horizontal and vertical coordinates represent the specific road endpoints; Representing specific road endpoints, i.e. individual nodes in the vertex set, each endpoint is defined by its geographic coordinates in space, serving as the basic spatial unit for network analysis; Let x represent the x-th road segment, which is a basic linear element constituting the complete road network; where x = 1, 2, ..., m, and m is the total number of road segments within the study area; R represents the set of the complete road network, which includes all road segments within the study area, i.e. E represents the set of edges in the road network, consisting of all continuous road segments. Edges are constructed based on the connectivity of the actual roads if and only if the two endpoints... and On the same road section A connecting edge is established between two adjacent elements only when they appear in the same element. Representing an edge The weights are the Euclidean distances between the two endpoints, calculated by the L2 norm of the coordinate difference. The result accurately reflects the actual road length. To sum the road segments within the study area; Used as a separator; It is a norm 2; S7.2: Locate the public service facility point to the nearest node in the road network. The calculation formula is as follows: In the formula, Indicates public service facilities The nearest node in the road network Represents road nodes and public service facilities The Euclidean distance between them; To select the node corresponding to the minimum function value; S7.3: Using the nearest facility point analysis tool, starting from the location of public service facilities, and with the service distance between the preprocessed age-group population distribution data and the living circle as the search radius, reachable nodes and reachable edges are extracted. The calculation formula is as follows: In the formula, j represents a public service facility; Indicates The graph centered on the target is a reachable subgraph. Let j be the set of reachable nodes for public service facility j; Let j be the set of reachable edges of public service facility j; Represents a node To the node The shortest path distance; For the living circle Suitable walking distance; S7.4: Perform Delaunay triangulation on the reachable node set and filter by the circumcircle radius threshold to generate the service boundary of public service facilities. This step transforms discrete reachable nodes into continuous service coverage areas, as shown in the following formula: In the formula, Rcircum represents the radius of the circumcircle of the Delaunay triangle; This is a preset threshold parameter used to control the compactness of the generated boundary. The corresponding radius threshold; S8: Based on the service scope of public service facilities and the demand level of the population of all ages, the overall supply and demand matching efficiency is obtained; S8 calculates the comprehensive supply-demand matching efficiency based on the service scope of public service facilities and the demand levels of the population across all age groups, including the following steps: S8.1: Based on the service scope of public service facilities, use spatial connectivity and summary statistics tools in ArcGIS to calculate the service coverage area of ​​various facilities within each living circle. This yields the public service facility supply quality coefficient, which reflects the adequacy of spatial coverage of facility services. A value closer to 1 indicates more adequacy of coverage. The formula is as follows: In the formula, For the living circle Internal public service facilities The supply quality coefficient; Indicates living circle Internal public service facilities The service area; Indicates living circle Internal public service facilities The total amount; Indicates living circle area; Indicates the social circle Sum all public service facilities j; S8.2: Public service facility location data are divided into five categories: commercial services, elderly care services, basic education, cultural and recreational activities, and medical and health services. The overall supply level is calculated by considering both the quantity and quality of facilities and services, and the supply quality coefficient of public service facilities is used as the basis for this calculation. This process yields the supply level of public service facilities. This step transforms the number of facilities into standardized population service capacity indicators, enabling horizontal comparability of supply levels across different categories of facilities and providing a quantitative basis for subsequent supply-demand matching analysis. The formula is as follows: In the formula, Indicates living circle Internal public service facilities The supply level, expressed in population, is measured in people; Indicates living circle Internal public service facilities The quantity, in units of locations; Indicates public service facilities The target for each thousand people is calculated per person / department. For the living circle Internal public service facilities The supply quality coefficient; S8.3: Based on the supply level of public service facilities, a comprehensive supply-demand matching index is obtained. This step assesses the relative deviation between the supply and demand of this type of facility. The formula is as follows: In the formula, For the living circle Chinese public service facilities The supply and demand matching index, its core is through comparison and The relative differences are used to assess the efficiency of supply and demand matching; and The sum is mainly used to standardize molecules, eliminate the influence of dimensions, and transform absolute differences into relative proportions, making the results comparable. For the living circle Internal public service facilities Supply level; For the living circle Internal public service facilities The demand level is the weighted average of the total age-based demand population; coefficient 2 mainly represents the supply and demand matching index. The range of values ​​for is expanded from [-1,1] to [-2,2] to more intuitively distinguish the degree of supply and demand imbalance; =0: Supply and demand are basically balanced and well matched; >0: Supply exceeds demand, resulting in redundancy; <0: Supply is less than demand, indicating a shortage; | The smaller the value, the higher the matching efficiency; S8.4: Based on the comprehensive supply and demand matching index, the comprehensive supply and demand matching efficiency is obtained, as shown in the following formula: In the formula, For the living circle The overall spatial supply and demand matching efficiency of all public service facilities; Indicates living circle Internal public service facilities The total amount; Life Circle Chinese public service facilities The supply and demand matching index; To express one's love for one's living circle Summation is performed on all public service facilities; The range of indicators is [ [2], the theoretical maximum value is 2, which means that the supply and demand of various facilities are perfectly matched; A higher value indicates a more reasonable facility configuration and higher matching efficiency; a lower value indicates a more serious facility mismatch, requiring key optimization.

[0023] The specific implementation method is as follows: Step 1: Collection and Preprocessing of Multi-Source Data. This stage requires collecting mobile phone signaling data, road network and water system data, population distribution data by age group, public service facility location data, and relevant planning standards for the study area. The content and types of collected data are shown in Table 1. Among them, mobile phone signaling data serves as the core data source, recording residents' spatial movement trajectories; road network and water system data provide physical evidence for delineating street boundaries; and facility POI data forms the basis for subsequent supply and demand analysis. All of these data need to undergo preprocessing steps such as coordinate unification, format standardization, and quality verification to ensure data consistency and reliability.

[0024] Table 1 Data Content and Type Step 2: Complete resident behavior feature extraction and spatial unit construction. Through in-depth processing of mobile phone signaling data, including data cleaning, trajectory reconstruction, and stop point identification, accurately extract residents' stable residence locations and their spatiotemporal characteristics. Simultaneously, based on physical boundary elements such as road networks, standardized street block units are identified and vectorized in the GIS platform. Stop points are associated with corresponding street blocks using spatial matching methods, thereby establishing a refined correspondence between resident activities and spatial units, forming a structure such as... Figure 2 The population distribution characteristics shown.

[0025] Step 3 involves constructing a dual-graph structure to represent complex spatial relationships. This step simultaneously establishes two complementary graph structures: one is a spatial topology graph, which, based on the geometric center coordinates of blocks, constructs spatial adjacency relationships through precise distance calculations to represent the physical proximity of blocks; the other is a pedestrian flow graph, which uses blocks as nodes and constructs connecting edges based on actual resident flow intensity to form a characteristic network reflecting functional connections. These two graph structures together form the basic framework for subsequent analysis.

[0026] Step 4: Perform graph embedding learning and community discovery. Graph convolutional network technology is used to learn the embeddings of the pedestrian flow graph, obtaining a low-dimensional vector representation of each block from a flow perspective. Then, the CESNA community discovery algorithm is called, fusing spatial topology and pedestrian flow feature embeddings to achieve intelligent clustering of block nodes. After spatial visualization processing, the final result is as follows: Figure 3 The results of the dynamic living circle division in the central urban area of ​​Kunming shown indicate that each identified community represents a living circle that matches the actual activity patterns of residents.

[0027] Step 5: Complete data preparation and parameter setting. Based on the identified living areas, conduct a refined assessment of facility supply and demand. After the living areas are identified, begin configuring parameters for all age groups and processing basic data. This stage requires collecting refined population data by age group and, based on the findings of behavioral geography, setting differentiated service radius parameters for different age groups. Simultaneously, establish a complete classification system for public service facilities (see Table 2), covering five major categories: commercial services, elderly care services, basic education, cultural and recreational activities, and medical and health services, and clarify the planning and configuration standards for each type of facility.

[0028] Table 2 Classification and Statistics of Public Service Facilities Step 6: Calculate the service range and supply level of facilities. Based on high-precision road network data, network analysis is used to generate the actual service range of each facility. By comprehensively considering multiple factors such as the number of facilities, service coverage quality, and planning standards, a standardized facility supply level is calculated. Spatial analysis is used to obtain the supply and demand matching index distribution of various facilities, forming a distribution such as... Figure 4 The spatial layout shown illustrates the matching of supply and demand for facilities.

[0029] Step 7: Calculate and assess the demand level across all age groups and the supply-demand matching. Construct a two-dimensional demand weight matrix for each age group and facility (see Table 3) to quantify the relative demand intensity for various facilities among different age groups. Use this matrix to weight the population data of each age group within the living area, calculating the weighted facility demand level across all ages. Finally, by comparing the supply level with the overall age-based demand level, calculate the supply-demand matching index and overall matching efficiency of individual facilities within the living area, accurately identifying areas of surplus and deficiency in facility allocation.

[0030] Table 3 Weights of Public Service Facility Demands for Different Age Groups This invention includes a dynamic identification module for urban living circles (such as...) Figure 6 The supply and demand matching assessment module includes a demand weight adjustment interface, as shown below. Figure 7 As shown, the interface for displaying the evaluation results of the supply and demand matching evaluation module is as follows: Figure 8 As shown, it supports interactive analysis and visual representation of planning decisions.

[0031] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for identifying residents' living circles based on their mobility and assessing the supply and demand of facilities suitable for all ages, characterized by: S1: Collect mobile phone signaling data, street data, road network and water system data, population distribution data by age group, public service facility location data and relevant planning standards in the study area and preprocess them respectively; S2: Extract resident stop points from the preprocessed mobile signaling data, associate resident stop points with the corresponding blocks, and extract the flow intensity between stop points within the blocks; S3: Perform spatial modeling of street data and flow intensity to construct a spatial topology map and pedestrian flow relationship map of the spatial modeling data; S4: Embedding learning is performed on the people flow relationship graph. The embedded people flow relationship graph and the spatial topology graph are then used to perform graph clustering using the CESNA algorithm. After multiple iterations, the living circle affiliation label is obtained. S5: Calculate the differentiated service distance for each living circle based on the preprocessed population distribution data by age group; S6: Construct the demand weight matrix of the preprocessed population distribution data by age group, and obtain the demand level of the population in all age groups based on the demand weight matrix. S7: Based on the preprocessed road network and water system data, public service facility location data, relevant planning standards, and preprocessed age-group population distribution data and service distance of the living circle, generate the service range of public service facilities; S8: Based on the service scope of public service facilities and the demand level of the population of all ages, the overall supply and demand matching efficiency is obtained.

2. The method for identifying living circles based on resident mobility and assessing the supply and demand of facilities suitable for all ages, as described in claim 1, is characterized in that: The S1 block data is obtained by identifying closed spatial units formed by the enclosed spatial elements in the urban road network and water system within the GIS platform, based on the spatial elements that have physical boundary functions. The data of these closed spatial units are then regarded as block data.

3. The method for identifying living circles based on resident mobility and assessing the supply and demand of facilities suitable for all ages, as described in claim 1, is characterized in that: The S3 space modeling includes the following steps: S3.1: Read the latitude and longitude coordinates of the geometric center point of the street block area layer generated in ArcGIS Pro; S3.2: Based on the preprocessed latitude and longitude coordinates of the block center points, the great circle distance between any two block center points is calculated using the Haversine formula. The specific calculation formula is as follows: In the formula, R represents the distance between the center points of two blocks; R is the Earth's radius. , These are the latitudes of the two points; , These are the longitudes of the two points; It is the arcsine function; It is a sine function; It is a cosine function; S3.3: Model using the geopy, networkx, and scikit-learn libraries to obtain the distance matrix.

4. The method for identifying living circles based on resident mobility and assessing the supply and demand of facilities suitable for all ages, as described in claim 3, is characterized in that: S3 constructs a spatial topology map and a pedestrian flow relationship map of the spatial modeling data; The spatial topology graph is constructed using the NearestNeighbors algorithm from the scikit-learn library based on the distance matrix. The pedestrian flow graph uses street blocks as nodes. Based on pedestrian flow data between blocks, the edge weights are calculated using the weighted average of weekday and weekend flow intensities to construct the topological structure of the pedestrian flow graph, which is stored as an adjacency matrix. ; The formula for calculating flow strength is: In the formula, This indicates the flow intensity from block a to block b; This indicates the frequency of movement from block a to block b; This represents the average distance traveled from block a to block b.

5. The method for identifying living circles based on resident mobility and assessing the supply and demand of facilities suitable for all ages, as described in claim 4, is characterized in that: S4 performs embedding learning on the people flow relationship graph, and then uses the CESNA algorithm to perform graph clustering on the embedded people flow relationship graph and the spatial topology graph. After multiple iterations, the living circle affiliation label is obtained, including the following steps: S4.1: Embedding learning is performed on the pedestrian flow graph using a graph convolutional network. During the propagation process of the graph convolutional network, the following formula is used for the hierarchical transfer of node features: In the formula, To represent the first in the network The node feature matrix of the layer; Indicates the first in the network The node feature matrix of the layer; For activation functions; Adjacency matrix The enhanced adjacency matrix formed after adding self-connections; yes The degree matrix; Indicates the first The trainable weight matrix of the layer; S4.2: Use the CESNA algorithm to perform graph clustering on the embedded pedestrian flow graph and spatial topology graph, randomly initialize the community membership matrix, and update the community membership matrix in each iteration. The community membership update formula of the CESNA algorithm is as follows: In the formula, Indicates the first One block; Indicates the first One community; Indicates a street Belongs to the community Membership degree; The membership component based on topology is obtained by calculating the structural similarity of blocks in the spatial topology graph using the networkx library. The membership component based on neighborhood attributes is obtained by calculating the cosine similarity between the neighborhood embedding vector and the community centroid of the embedded pedestrian flow graph using the NumPy library. S4.3: After a fixed number of iterations and convergence, determine the final living circle affiliation label of each block based on its maximum community affiliation degree; In the formula, Indicates the first One block; Indicates the first A living circle; For the first The final living circle affiliation label of each neighborhood; To represent taking the option that maximizes the function value value; After a fixed number of iterations, the th... The block belongs to the first Community affiliation of each living circle.

6. The method for identifying living circles based on resident mobility and assessing the supply and demand of facilities suitable for all ages, as described in claim 1, is characterized in that: The formula used by S5 to calculate the service distance between the preprocessed age-group population distribution data and the living circle is as follows: In the formula, For the living circle Suitable walking distance; , , , respectively living circle The proportions of young people, middle-aged people, teenagers and the elderly.

7. The method for identifying living circles based on resident mobility and assessing the supply and demand of facilities suitable for all ages, as described in claim 1, is characterized in that: S6 constructs a demand weight matrix for the preprocessed age-group population distribution data, and obtains the demand level for all age groups based on the demand weight matrix, including the following steps: S6.1: The formula for the demand weight matrix of the preprocessed age-group population distribution data is as follows: In the formula, express Age groups' demand for public service facilities Demand weight, express Public service facilities for different age groups The total number of people in the grid; Indicates the study area Total population of the same age group; S6.2: Construct a two-dimensional demand weight matrix for age groups and facilities based on the demand weight matrix formula; S6.3: Obtain the population demand level for all age groups based on the demand weight matrix, using the following formula: in, Indicates living circle Internal public service facilities The total age-weighted demand level of the population; Indicates living circle Inside Population size of the relevant age group; express Age groups' demand for public service facilities Demand weight; To sum the information from the four age groups.

8. The method for identifying living circles based on resident mobility and assessing the supply and demand of facilities suitable for all ages, as described in claim 6, is characterized in that: S7 generates the service area of ​​public service facilities based on preprocessed road network and water system data, public service facility data, relevant planning standards, and preprocessed age-group population distribution data and service distance of living circles, including the following steps: S7.1: In ArcGIS Pro, use the Network Analyst module to construct a road network dataset, abstracting roads into a network model containing nodes and edges, as shown in the following formula: In the formula, G represents the road network diagram; V represents a node of the road network, which consists of the coordinates of all road endpoints; and The horizontal and vertical coordinates represent the specific road endpoints; Represents a specific road endpoint; R represents the x-th road segment; R represents the set of the complete road network, which includes all road segments within the study area; E represents the set of edges of the road network, which consists of all continuous road segments. Representing an edge The weights are the Euclidean distances between the two endpoints, calculated by the L2 norm of the coordinate difference. get; To sum the road segments within the study area; Used as a separator; It is a norm 2; S7.2: Locate the facility point to the nearest node in the road network. The calculation formula is as follows: In the formula, Indicates public service facilities The nearest node in the road network Represents road nodes and public service facilities The Euclidean distance between them; To select the node corresponding to the minimum function value; S7.3: Using the nearest facility point analysis tool, starting from the location of public service facilities, and with the service distance between the preprocessed age-group population distribution data and the living circle as the search radius, reachable nodes and reachable edges are extracted. The calculation formula is as follows: In the formula, j represents a public service facility; Indicates The graph centered on the target is a reachable subgraph. Let j be the set of reachable nodes for public service facility j; Let j be the set of reachable edges of public service facility j; Represents a node To the node The shortest path distance; For the living circle Suitable walking distance; S7.4: Perform Delaunay triangulation on the reachable node set and filter by the circumcircle radius threshold to generate the service boundary of public service facilities. The calculation formula is as follows: In the formula, This represents the radius of the circumcircle of the Delaunay triangle; This is a preset threshold parameter; This is the corresponding radius threshold.

9. The method for identifying living circles based on resident mobility and assessing the supply and demand of facilities suitable for all ages, as described in claim 8, is characterized in that: The S8 method, based on the service scope of public service facilities and the demand levels of the entire age group, obtains the comprehensive supply and demand matching efficiency, including the following steps: S8.1: Based on the service scope of public service facilities, use spatial connectivity and summary statistics tools in ArcGIS to calculate the service coverage area of ​​various facilities within each living circle. The quality coefficient of public service facility supply is obtained using the following formula: In the formula, For the living circle Internal public service facilities The supply quality coefficient; Indicates living circle Internal public service facilities The service area; Indicates living circle Internal public service facilities The total amount; Indicates living circle area; Indicates the social circle Sum all public service facilities j; S8.2: The supply level of public service facilities is obtained based on the quality coefficient of public service facility supply, using the following formula: In the formula, Indicates living circle Internal public service facilities Supply level; Indicates living circle Internal public service facilities Quantity; Indicates public service facilities The plan targets a thousand-person quota; For the living circle Internal public service facilities The supply quality coefficient; S8.3: Based on the supply level of public service facilities, a comprehensive supply-demand matching index is obtained, as shown in the following formula: In the formula, For the living circle Chinese public service facilities The supply and demand matching index; For the living circle Internal public service facilities Supply level; For the living circle Internal public service facilities The total age-weighted demand level of the population; S8.4: Based on the comprehensive supply and demand matching index, the comprehensive supply and demand matching efficiency is obtained, as shown in the following formula: In the formula, For the living circle The overall spatial supply and demand matching efficiency of all public service facilities; Indicates living circle Internal public service facilities The total amount; Life Circle Chinese public service facilities The supply and demand matching index; To express one's love for one's living circle Summation of all public service facilities It represents the absolute value.

Images