Urban night vitality driving factor evaluation method

Abstract

The invention provides an evaluation method for urban night vitality driving factors. Comprising the following steps: (1) collecting and processing multi-source heterogeneous spatio-temporal data, and constructing a multi-level self-adaptive grid by a dynamic recursive algorithm; (2) constructing a multi-dimensional evaluation system, and establishing an explicit-hidden variable mapping relationship between a measurement model and a structure model; (3) establishing a space structure equation model, and quantifying direct effects and space overflow effects of natural, social and economic factors on urban night vitality; (4) carrying out model verification and sensitivity analysis, and evaluating model performance by adopting goodness of fit, standardized factor load and the like; and (5) evaluating urban night vitality driving factors, and outputting a high-resolution urban night vitality index thermodynamic diagram and influence factor contribution degrees. According to the method, spatial correlation is introduced into the structural equation model, spatial heterogeneity characteristics of night vitality distribution are considered, core driving factors influencing urban night vitality and an action mechanism of the core driving factors are quantified, and a scientific basis is provided for urban planning and policy making.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of urban research; in particular, it relates to a method for evaluating driving factors of urban night vitality. BACKGROUND

[0002] With the development of cities entering a new stage, urban night vitality has become a key indicator to measure the economic prosperity, cultural attractiveness and quality of life of residents. Night economy plays a significant role in promoting consumption, creating employment and enhancing the image of the city, so its driving factors and influence paths have attracted widespread attention from academia. However, urban night vitality is a complex system phenomenon driven by multiple dimensions and multiple scales, and there are often interactions and spatiotemporal dynamic changes among these factors, making it difficult to accurately identify the core driving factors, quantify their contribution and clarify their influence paths.

[0003] In rapidly developing cities, due to differences in economic development level, built environment characteristics, natural environment and living conditions, urban night vitality shows obvious spatial heterogeneity, making it challenging to fully capture and understand the driving mechanism. Although massive mobile positioning data, social media data, and light remote sensing data provide new perspectives and new data sources for research, existing methods still struggle to effectively integrate these multi-source heterogeneous data to build a comprehensive evaluation framework that can fully depict the formation mechanism of night vitality and analyze the interaction and influence paths of multiple driving factors. This makes it difficult for city managers to accurately identify the core elements of improving night vitality in specific areas, resulting in insufficient targeting of resource allocation and policy making. Therefore, how to effectively integrate multi-source heterogeneous data, fully consider spatial dependence, and explore and quantify the core driving factors and their mechanisms that affect urban night vitality from multiple dimensions and multiple scales has become an important issue in current urban research and management practice. SUMMARY

[0004] The purpose of the present application is to provide a method for evaluating driving factors of urban night vitality, which solves the problems existing in the background art and realizes the evaluation of urban night vitality influencing factors at multiple levels of grid scale.

[0005] The present application is realized by the following technical solutions:

[0006] The present application relates to a method for evaluating driving factors of urban night vitality, comprising the following steps:

[0007] Step 1, multi-source heterogeneous urban spatiotemporal data acquisition and processing;

[0008] Step 2, multi-dimensional evaluation system construction;

[0009] Step 3, spatial structure equation model establishment;

[0010] Step 4, spatial structure equation model verification and sensitivity analysis;

[0011] Step 5, evaluation of driving factors of urban night vitality.

[0012] Preferably, in step 1, the data collection method is specifically: collecting multi-source data related to urban night vitality through various ways such as public data opening platform, government official data platform, remote sensing satellite data service, commercial geographic information service, etc. The data sources include government statistical data, urban infrastructure data, remote sensing data, social media data, etc., specifically including night light, population density, interest point, normalized vegetation index, elevation, per capita gross national product, bus station, road network, PM2.5, industrial and commercial enterprise, building contour, etc.; data processing includes steps such as removing missing values, abnormal value processing, and data standardization and normalization, to ensure the consistency and quality of the data.

[0013] Preferably, in step 1, the processing is specifically: multi-level grid division processing is performed on the collected data.

[0014] Urban night vitality often shows different distribution patterns in the study area, for example, the suburban population gathering area is scattered, there are large areas with few people, and in the densely populated urban area, urban vitality often shows spatial agglomeration characteristics. With the same resolution, there may be data redundancy in some suburbs, and it is more difficult to find low-activity areas in high-activity areas in the city. Multi-level grid division can effectively identify hotspots of urban night vitality, ensuring analysis accuracy while improving computing efficiency.

[0015] With the boundary of the study area as the range, set the grid division threshold α, the initial spatial resolution γ s and the terminal spatial resolution γ e , the threshold can be set by POI density, road network density, etc. The study area is divided into grids according to the initial spatial resolution γ s , and the divided grids are judged in turn. If the grid density is greater than the threshold α or the grid resolution is less than γ e , further divide with γ s / 2 as the resolution, recursively execute the above steps, until all grid densities are less than α or resolutions are greater than or equal to γ e . Among them, the initial spatial resolution, the terminal spatial resolution and the road network threshold can be determined according to the size of the study area and the research target. The divided grid can provide spatial units for subsequent analysis.

[0016] Preferably, in step 2, the multi-dimensional evaluation system construction includes: construction of latent variables, selection of observed variables, and standardization of observed variables.

[0017] Preferably, the potential variables include four potential variables of urban night vitality f1, natural environment f2, social environment f3, and economic development f4, wherein the natural environment, the social environment, and the economic development are main factors affecting the urban night vitality. The potential variables can be determined according to existing literatures, expert consultation, or a multivariate analysis method.

[0018] The structural equation model is a model capable of modeling potential variables by using observation variables, thereby helping to understand the mutual relationship between various components in a complex system. In addition, the model can construct direct effects and indirect effects, thereby revealing complex relationships between multiple variables. The urban night vitality of a region is usually closely related to factors such as natural, social, and economic development of the region;

[0019] The observation variables are usually directly measurable data and can be used to quantify the potential variables that cannot be directly measured. The observation variables are obtained by means of data mining, statistical investigation, or remote sensing monitoring. According to the four potential variables of urban night vitality, natural environment, social environment, and economic development, the following observation variables are selected: (1) the observation variables of urban night vitality include night light value and annual change rate of night light; (2) the observation variables of natural environment include slope, PM2.5, and normalized vegetation index; (3) the observation variables of social environment include volume rate, bus station density, and road network density; (4) the observation variables of economic development include industrial and commercial enterprise density, proportion of tertiary industry, and per capita GDP; the observation variables are determined according to the research region and the availability of data. Among them, the night light intensity directly reflects the distribution of artificial light sources in the region, and existing studies have shown that it is closely related to economic activities (such as GDP and per capita income). At the same time, the resolution of night light data reaches 500 meters and the coverage is wide, which can identify the differences within the city. The data directly representing the urban night vitality are mostly statistical units based on administrative division (such as population) or difficult to obtain (such as mobile positioning data), and the statistical scale is large and the coverage is not comprehensive. Therefore, the night light as a proxy indicator of urban night vitality is reasonable.

[0020] The calculation methods of the observation variables are as follows:

[0021] Night light value NTL: the average value of night light in the region;

[0022] Night light change rate NTLC: the growth rate of night light value in the region from 2015 to 2024;

[0023] Slope: average slope in the region;

[0024] PM2.5: average PM2.5 in the region;

[0025] Normalized vegetation index NDVI: average NDVI value in the region;

[0026] FAR: FAR is the floor area ratio of buildings in the region, which is equal to the total building area / area of the region;

[0027] BSD: BSD is the number of bus stops in the region / area of the region;

[0028] RD: RD is the total length of the road network in the region / area of the region;

[0029] BD: BD is the number of business enterprises in the region / area of the region;

[0030] TIP: TIP is the proportion of the tertiary industry in the region, which is equal to the number of tertiary industry enterprises in the region / number of all enterprises in the region;

[0031] GDP per capita: GDP per capita in the region;

[0032] The standardization of the observation variables is as follows: Z-score standardization (mean is 0, variance is 1) is adopted.

[0033]

[0034] Wherein, x is the original data, μ and σ are the mean and standard deviation of the original data respectively; if the Z-score is positive, it means that the data point is higher than the mean; if the Z-score is negative, it means that the data point is lower than the mean; the greater the absolute value of the Z-score, the farther the data point is from the mean. After standardization, all variables have the same dimension, which helps the stability and effectiveness of model estimation.

[0035] Preferably, in step 3, the spatial structure equation model is composed of a measurement model (outer model) and a structure model (inner model). The measurement model defines the relationship between the observation variables and the latent variables, and the structure model defines the relationship between the latent variables. The structure model involved in the present application introduces a linear synergistic model to model the spatial correlation;

[0036] Preferably, the measurement model is used to verify whether the latent variable can be effectively represented by the corresponding observation variable; the present application evaluates the contribution of each observation variable to the latent variable by constructing the path relationship between the latent variable and its observation variable, and then determines the weight of each urban night-time vitality factor. The formula is as follows:

[0037] θ=λ θ f+ε

[0038] Wherein, f represents the latent variable; θ represents the observation variable; λ θ represents the factor loading between the latent variable and the observation variable, which represents the correlation between the two; ε represents the measurement error;

[0039] The structural model is used to reveal the causal path between each potential variable, analyze the correlation between potential variables, and quantify the influence of each factor on the urban night vitality. In the present application, the structural model explores how the factors such as economic development, natural environment and social environment interact with each other, and then affect the level of urban night vitality.

[0040] For potential variables, the calculation formula of the structural model is:

[0041] f1=γ1f2+γ2f3+γ3f4+δ

[0042] Wherein, γ1, γ2, γ3 are unknown constants, f1, f2, f3, f4 represent potential variables, which are urban night vitality, natural environment, social environment and economic development respectively; it is assumed that δ has a univariate conditional autoregressive (CAR) covariance structure, and the covariance parameter is a δ =(τ δ ,ρ δ ), wherein τ δ is a precision parameter, and ρ δ is a spatial correlation parameter;

[0043] The posterior calculation uses Markov chain Monte Carlo (MCMC) method, which can randomly sample from the marginal posterior distribution of each parameter and potential factor, so as to obtain the posterior summary such as the posterior mean and quantile of the parameter.

[0044] The traditional structural equation model assumes that the potential factors are not correlated between regions, but in the evaluation of urban night vitality, adjacent or close regions may have similar vitality values or natural, social and economic environment, that is, they show certain spatial correlation;

[0045] The spatial factor model is introduced into the structural equation model, specifically: considering the spatial correlation between potential variables. Through spatial correlation modeling, the spatial agglomeration effect of urban night vitality can be revealed, and the influence of spatial factors on vitality evaluation can be analyzed.

[0046] The joint distribution of (f2, f3, f4) is modeled using a linear co-integration model: T

[0047]

[0048] Wherein, v1, v2, v3 represent the vectors of independent spatial processes in the study area; it is assumed that v1, v2, v3 have a CAR covariance structure, and the overall scale parameter is set to 1, and the spatial correlation parameters are and

[0049] ​A1, A4, A7 represent the contribution to v1, A2, A5, A8 represent the contribution to v2, and A3, A6, A9 represent the contribution to v3.

[0050] Preferably, in step 4, the spatial structure equation model verification and sensitivity analysis comprises: model calculation, parameter analysis and model debugging.

[0051] Preferably, the model calculation specifically is:

[0052] The R language and Openbugs software are used to import the calculated observation variables and neighborhood matrix into the software to realize compilation and build a preliminary structure equation model. According to the standardized path coefficients between the latent variables in the Openbugs software, the influence relationship of the natural environment, social environment and economic conditions on the urban night vitality is obtained, and the urban night vitality index is obtained based on NTL and NTLD;

[0053] Specific steps are: the R language and Openbugs software are used to complete model writing, the observation variables and spatial neighborhood matrix are imported into the Openbugs software as input data, the standardized path coefficients between the latent variables and the urban night vitality index can be obtained after the model is run, the influence relationship of the natural environment, social environment and economic conditions on the urban night vitality is analyzed according to the path coefficients, and the urban night vitality influence factor evaluation model is built;

[0054] Wherein, the spatial neighborhood matrix is specifically defined as matrix If The matrix element is defined as Otherwise w ij = 0, wherein n represents the number of grids, is the distance between grid i and j, and theta is the distance threshold value;

[0055] The parameter analysis specifically is: according to the spatial structure equation model calculation evaluation index, the GFI value needs to be greater than 0.9, the factor loading of the latent variable and the observation variable needs to be significant and the absolute value needs to be greater than 0.5, and the spatial autocorrelation of the model residual should not be significant. According to the above index, the model running logic is adjusted, the path that does not reach the significant level is deleted, the variable definition is optimized or the spatial weight matrix is adjusted, and finally the urban night vitality evaluation model is obtained, so as to ensure the reliability of the model in statistics and theory.

[0056] The model debugging is specifically: according to the results of model evaluation, necessary model correction is carried out. The influence path between variables in the model is judged, if the influence path is not significant, it is removed, if the goodness of fit of the model does not meet the requirements, the path coefficient and the relationship between potential variables need to be adjusted, the model correction also includes optimizing the selection of observation variables, adjusting the measurement error and eliminating the multicollinearity and other problems. Through repeated correction, it is ensured that the model can accurately reflect the influencing factors and spatial distribution characteristics of urban night vitality. In the evaluation process, the necessity of spatial factors can also be verified by comparing other models (such as traditional structural equation model).

[0057] Preferably, in step 5, the urban night vitality driving factor evaluation is specifically: after the model evaluation and correction are completed, the analysis of urban night vitality and each influencing factor is carried out, according to the results of the spatial structural equation model, the influence of economic development, natural environment and social environment on urban night vitality is quantified, and the spatial effect is analyzed. Not only helps to understand the multi-dimensional causes of night vitality, but also reveals the spatial differences of night vitality in different regions. According to the analysis results, scientific basis is provided for the improvement of urban night vitality and urban planning.

[0058] The storage medium used in the application is used to store computer executable instructions, which execute the above-mentioned urban block vitality evaluation method based on the spatial projection pursuit model when executed by the processor.

[0059] In summary, the application uses multi-source heterogeneous data to obtain observation variables and potential variables affecting urban night vitality, constructs a multi-level grid as an evaluation unit, proposes a spatial structural equation model considering spatial dependence, and evaluates the driving factors and influence paths of urban night vitality. The method defines four potential variables and 12 observation variables, considers the spatial correlation and influence relationship between potential variables, and obtains the influence degree of natural environment, social environment and economic development on urban night vitality. The application can better quantitatively evaluate the influencing factors of urban night vitality, explore and quantify the core driving factors and their action mechanisms affecting urban night vitality from multiple dimensions and multiple scales, and promote the high-quality development of cities.

[0060] The application has the following beneficial effects:

[0061] (1) A multi-level grid is constructed as an evaluation unit of urban night vitality, which can effectively identify the hot spot area of urban night vitality, ensure the analysis accuracy and improve the calculation efficiency.

[0062] (2) The traditional structural equation model is improved, the spatial correlation between variables is considered, and the accuracy and reliability of urban night vitality evaluation are improved.

[0063] (3) The spatial structure equation model considers the direct and indirect effects between urban night vitality and natural, social, economic and other influencing factors, thereby revealing the complex relationship between multiple variables, and providing a scientific basis for urban planning and policy making. BRIEF DESCRIPTION OF DRAWINGS

[0064] Figure 1 is a flowchart of the urban night vitality driving factor evaluation method involved in the present application;

[0065] Figure 2 is a spatial structure equation model architecture diagram involved in the present application;

[0066] Figure 3 is a multi-level grid map of the six urban districts of Xi'an involved in the present application. DETAILED DESCRIPTION

[0067] The present application will be described in detail below with specific embodiments. It should be noted that the following implementation examples are only further illustrations of the present application, and the protection scope of the present application is not limited to the following examples.

[0068] EMBODIMENT

[0069] The urban night vitality driving factor evaluation method of the present embodiment, as shown in Figure 1 , includes the following steps:

[0070] Step 1, multi-source heterogeneous urban spatio-temporal data collection and processing;

[0071] The data collection specifically includes collecting multi-source data related to urban night vitality through various ways such as public data opening platform, government official data platform, remote sensing satellite data service, commercial geographic information service, etc. The data sources include government statistical data, urban infrastructure data, remote sensing data, social media data, etc.; among them, the key collected data includes night light, population density, interest point, normalized vegetation index, elevation, per capita gross national product, bus station, road network, PM2.5, industrial and commercial enterprise, building contour, etc., and the data name, year, source and data type are shown in Table 1. The data processing includes steps such as removing missing values, abnormal value processing, data standardization and normalization, etc., to ensure the consistency and quality of the data.

[0072] Table 1 Data Sources

[0073] Data name Year Source Type Night light 2023 NPP / VIIRS Grid Population density 2020 WorldPop Grid Point of interest 2024 Gaode map Vector Normalized difference vegetation index 2022 National ecological science data center Grid Elevation 2022 ASF ALSO Grid GDP per capita 2020 Resource and environment science data registration and publishing system Grid Bus station 2024 Gaode map Vector Road network 2024 OpenStreetMap Vector PM2.5 2023 Zenodo Grid Industrial and commercial enterprise 2022 Market supervision administration Vector Building contour 2020 Zenodo Vector

[0074] The processing specifically includes multi-level grid division processing of the collected data;

[0075] The distribution of urban night-time vitality in the study area is different, for example, the suburban population gathering area is scattered, and there are large areas with few people. In the densely populated urban area, the urban vitality often shows spatial agglomeration. Using the same resolution, there will be data redundancy in some suburbs, and it is more difficult to find low-activity areas in high-activity areas in the city. Multi-level grid division can effectively identify hotspots of urban night-time vitality, ensuring analysis accuracy while improving computing efficiency.

[0076] Taking the six districts of Xi'an City as the research area, setting the grid division threshold α = 20, the initial spatial resolution γ s = 4km×4km and the final spatial resolution γ e = 0.02km×0.02km, using the POI density in the grid to define the judgment index. The study area is divided into grids according to the initial spatial resolution γ s , as shown in Figure 3 , the divided grids are judged in turn. If the POI density in the grid is greater than the threshold α or the grid resolution is less than γ e , further divide it with γ s / 2 as the resolution, recursively execute the above steps until all grids have POI density less than α or resolution greater than or equal to γ e , Figure 3 The multi-level grid division results are expressed in

[0077] Step 2, construction of multi-dimensional evaluation system;

[0078] The construction of the evaluation system includes: the construction of potential variables, the selection of observation variables, and the standardization of observation variables;

[0079] The potential variables include urban night-time vitality f1, natural environment f2, social environment f3, and economic development f4. Natural environment, social environment, and economic development are the main factors affecting urban night-time vitality. Potential variables can be determined according to existing literature, expert consultation, or multivariate analysis methods.

[0080] Structural equation modeling is a model that can use observation variables to model potential variables, helping to understand the relationship between components in complex systems. In addition, the model can construct direct and indirect effects, revealing the complex relationship between multiple variables. The urban night-time vitality of an area is usually closely related to factors such as natural, social, and economic development in that area;

[0081] The observation variables are usually directly measurable data that can be used to quantify latent variables that cannot be directly measured. The observation variables are obtained through various means such as data mining, statistical investigation, or remote sensing monitoring. According to the four latent variables of urban night vitality, natural environment, social environment, and economic development, the following observation variables are selected: (1) the observation variables of urban night vitality include night light value and annual change rate of night light; (2) the observation variables of natural environment include slope, PM2.5, and normalized vegetation index; (3) the observation variables of social environment include FAR, bus station density, and road network density; (4) the observation variables of economic development include business density, proportion of tertiary industry, and per capita GDP; and the observation variables are determined according to the research region and the availability of data.

[0082] The calculation methods of the observation variables are as follows:

[0083] Night light value NTL: the average value of night light in the region;

[0084] Night light change rate NTLC: the growth rate of night light value in the region from 2015 to 2024;

[0085] Slope: the average slope in the region;

[0086] PM2.5: the average PM2.5 in the region;

[0087] Normalized vegetation index NDVI: the average NDVI value in the region;

[0088] FAR: the FAR of the region, which is equal to the total building area / region area;

[0089] Bus station density BSD: the number of bus stations in the region / region area;

[0090] Road network density RD: the total length of road network in the region / region area;

[0091] Business density BD: the number of businesses in the region / region area;

[0092] Proportion of tertiary industry TIP: the number of tertiary industry enterprises in the region / total number of enterprises in the region;

[0093] Per capita GDP: per capita GDP in the region;

[0094] The standardization of the observation variables is specifically: using the standard score Z-score for standardization (mean is 0, variance is 1).

[0095]

[0096] Where x is the original data, μ and σ are the mean and standard deviation of the original data, respectively. If the Z-score is positive, it means that the data point is higher than the mean. If the Z-score is negative, it means that the data point is lower than the mean. The greater the absolute value of the Z-score, the farther the data point is from the mean. After standardization, all variables have the same dimension, which helps to improve the stability and effectiveness of model estimation.

[0097] Step 3, the spatial structure equation model is established, as shown in Figure 2 ;

[0098] In step 3, the spatial structure equation model is composed of a measurement model (outer model) and a structure model (inner model). The measurement model defines the relationship between observed variables and latent variables, and the structure model defines the relationship between latent variables. The structure model introduced in the present application models spatial correlation by introducing a linear synergistic model. The measurement model is used to verify whether the latent variables can be effectively represented by the corresponding observed variables, as shown in all arrows in the figure except the dashed box. By constructing the path relationship between the latent variables and their observed variables, the present application evaluates the contribution of each observed variable to the latent variables, and then determines the weight of each city night vitality driving factor. The formula is as follows: Figure 2

[0099] θ=λ θ f+ε

[0100] Where f represents the latent variable, θ represents the observed variable, λ represents the factor loading between the latent variable and the observed variable, which represents the correlation between the two, and ε represents the measurement error. θ

[0101] The structure model is used to reveal the causal path between latent variables, analyze the correlation between latent variables, and quantify the influence of each factor on city night vitality, as shown in all arrow representations in the figure except the arrows contained in the dashed box. In this patent, the structure model explores how economic development, natural environment, social environment and other factors interact with each other and then affect the level of city night vitality. Figure 2

[0102] For latent variables, the calculation formula of the structure model is as follows:

[0103] f1=γ1f2+γ2f3+γ3f4+δ

[0104] Where γ1, γ2, and γ3 are unknown constants, f1, f2, f3, and f4 represent latent variables, and δ is assumed to have a univariate conditional autoregressive (CAR) covariance structure with covariance parameters a δ =(τ δ ,ρ δ ), where τ δ ​​​is the precision parameter, p δ is the spatial correlation parameter;

[0105] Posterior computation uses Markov Chain Monte Carlo (MCMC) methods, which are able to sample from the marginal posterior distribution of each parameter and latent factor, and thus obtain posterior summaries such as the posterior mean and quantiles of the parameters.

[0106] Traditional structural equation models assume that latent factors are uncorrelated across regions, but in the evaluation of urban night-time vitality, adjacent or close regions may have similar night-time vitality indexes or natural, social, and economic environments, i.e., exhibit certain spatial correlation.

[0107] The spatial factor model is introduced into the structural equation model, specifically, the spatial correlation between latent variables is considered. Through spatial correlation modeling, the spatial agglomeration effect of urban night-time vitality can be revealed, and the influence of spatial factors on the evaluation of night-time vitality can be analyzed.

[0108] The joint distribution of (f2, f3, f4) is modeled using a linear copula: T

[0109]

[0110] where v1, v2, v3 represent vectors of independent spatial processes in the study area; it is assumed that v1, v2, v3 have a CAR covariance structure, and the overall scale parameter is set to 1, and the spatial correlation parameters are and

[0111] is the coefficient matrix, which defines how the latent variables are generated from the spatial processes, each row represents the contribution of a spatial process to different latent variables, a1, a4, a7 represent the contribution to v1, a2, a5, a8 represent the contribution to v2, and a3, a6, a9 represent the contribution to v3.

[0112] Step 4, spatial structural equation model verification and sensitivity analysis;

[0113] Spatial structural equation model verification and sensitivity analysis includes model calculation, parameter analysis, and model debugging.

[0114] wherein the model calculation is specifically:

[0115] ​The R language and Openbugs software are used to import the calculated observation variables and neighborhood matrix into the software to realize compilation and build a preliminary structural equation model. According to the standardized path coefficients between the potential variables in the Openbugs software, the influence relationship of the natural environment, social environment and economic conditions on the urban night vitality is obtained, and the urban night vitality index is obtained based on NTL and NTLD;

[0116] The specific steps are: using R language and Openbugs software to complete model writing, importing the observation variables and spatial neighborhood matrix into the Openbugs software as input data, and obtaining the standardized path coefficients between the potential variables and the urban night vitality index after the model running, and analyzing the influence relationship of the natural environment, social environment and economic conditions on the urban night vitality according to the path coefficients, and building an urban night vitality influence factor evaluation model.

[0117] The spatial neighborhood matrix is specifically defined as a matrix If The matrix element is defined as Otherwise w ij = 0, wherein n represents the number of grids, is the distance between grids i and j, and θ is the distance threshold value.

[0118] The parameter analysis is specifically: Figure 2 The spatial structural equation model measurement results of the six districts of Xi'an city as the research area are shown in the table, wherein the GFI value is greater than 0.9, the factor load of the potential variables and the observation variables is greater than 0.5 except PM2.5, and the spatial autocorrelation of the model residual is not significant, which indicates that the observation variables have a strong contribution to the potential variables. From the path coefficient, the natural environment has a negative influence on the urban night vitality, and the social environment and economic environment have a positive influence on the urban night vitality, which indicates that the urban night vitality is lower in the areas located in the inconvenient transportation, underdeveloped economy and relatively poor natural environment, which indicates that the model has certain reliability and interpretability.

[0119] The model debugging is specifically: judging the influence path between the variables in the model, wherein the PM2.5 influence path is not significant and should be removed; adjusting the path coefficient and the relationship between the potential variables to ensure that the fitting degree of the model meets the requirements; through repeated correction, the model can accurately reflect the influence factors and spatial distribution characteristics of the urban night vitality. In the evaluation process, the necessity of the spatial factors is verified by comparing the traditional structural equation model.

[0120] The model running results show that PM2.5 does not show statistical significance, which may be that the variable cannot effectively explain the potential variables or there is collinearity with other variables, and it is not suitable for measuring the potential variable. After the model adjustment, the variable will be deleted, soFigure 2 The variable is not included in the middle.

[0121] Step 5, evaluation of urban night vitality driving factors; after the model evaluation and correction are completed, the analysis of urban night vitality and various influence factors is carried out; according to the results of the spatial structure equation model, the influence of economic development, natural environment and social environment on urban night vitality is quantified, and the spatial effect is analyzed. Not only can it help to understand the multi-dimensional causes of night vitality, but also can reveal the spatial differences of night vitality degree in different regions. According to the analysis results, scientific basis is provided for the improvement of urban night vitality and urban planning.

[0122] In summary, the present application uses multi-source heterogeneous data to obtain observation variables and potential variables affecting urban night vitality, constructs a multi-level grid as an evaluation unit, and proposes a spatial structure equation model considering spatial dependence to evaluate the driving factors and influence paths of urban night vitality. The method defines four potential variables and 12 observation variables, considers the spatial correlation and influence relationship between potential variables, and obtains the influence degree of natural environment, social environment and economic development on urban night vitality. The present application can better quantitatively evaluate the factors affecting urban night vitality, explore and quantify the core driving factors and their action mechanisms affecting urban night vitality from multiple dimensions and multiple scales, and promote the high-quality development of cities.

[0123] Obviously, those skilled in the art should understand that the units or steps of the present application described above can be realized by general computing devices, which can be concentrated on a single computing device, and optionally, they can be realized by computer device executable program codes, so that they can be stored in storage devices and executed by computing devices, or they can be manufactured into individual integrated circuit modules, or a plurality of modules or steps in them can be manufactured into a single integrated circuit module to realize. Thus, the present application is not limited to any specific combination of hardware and software.

[0124] The specific embodiments of the present application are described above. It should be understood that the present application is not limited to the above specific embodiments, and those skilled in the art can make various modifications or changes within the scope of the claims, which does not affect the essence of the present application.

Claims

1. A method for evaluating the driving factors of urban nighttime vitality, characterized in that: The following steps are involved: Step 1: Collection and processing of multi-source heterogeneous urban spatiotemporal data; Step 2: Construction of a multi-dimensional evaluation system; Step 3, establishing the spatial structural equation model; Step 4: Spatial structural equation model verification and sensitivity analysis; Step 5: Evaluate the driving factors of urban nighttime vitality.

2. The method for evaluating driving factors of urban nighttime vitality according to claim 1, characterized in that: In step 1, the data collection method is specifically: collecting multi-source data related to the city's nighttime vitality through various channels such as public data open platforms, government data platforms, remote sensing satellite data services, and commercial geographic information services; the data includes: night lights, population density, points of interest, normalized vegetation index, elevation, per capita GDP, bus stops, road network, PM2.5, industrial and commercial enterprises, and building outlines.

3. The method for evaluating driving factors of urban nighttime vitality according to claim 1, characterized in that: In step 1, the processing is specifically: multi-level grid division processing is performed on the collected data, specifically: Taking the boundary of the study area as the range, set the grid division threshold α and the initial spatial resolution γ s and the final spatial resolution γ e , the study area is divided into two parts according to the initial spatial resolution γ s Grid division is performed, and the divided grids are judged in turn. If the grid judgment index is greater than the threshold α or the grid resolution is less than γ e , with γ s / 2 is the resolution for further division, and the above steps are recursively performed until the road network density in all grids is less than α or the resolution is greater than or equal to γ e until.

4. The method for evaluating driving factors of urban nighttime vitality according to claim 1, characterized in that: In step 2, the construction of the multidimensional evaluation system includes: construction of latent variables, selection of observed variables, and standardization of observed variables.

5. The method for evaluating driving factors of urban nighttime vitality according to claim 4, characterized in that: The latent variables include four latent variables: urban nighttime vitality f1, natural environment f2, social environment f3, and economic development f4. Among them, natural environment, social environment, and economic development are the main factors affecting urban nighttime vitality; The observation variables are selected based on the four potential variables of urban nighttime vitality, natural environment, social environment, and economic development: (1) The observation variables of urban nighttime vitality include nighttime light value and annual change rate of nighttime light; (2) The observation variables of natural environment include slope, PM2.5, and normalized vegetation index; (3) The observation variables of social environment include floor area ratio, bus station density, and road network density; (4) The observation variables of economic development include industrial and commercial enterprise density, proportion of tertiary industry, and GDP per capita; Deterministic observation variables are obtained based on the research area and data; The calculation method of each observed variable is as follows: Night light value NTL: the average night light value in the area; Night Light Change Rate (NTLC): The growth rate of night light values ​​in the region from 2015 to 2024; Slope: average slope in the area; PM2.5: Average PM2.5 in the area; Normalized Difference Vegetation Index NDVI: the average NDVI value in the region; Floor Area Ratio FAR: Floor area ratio of buildings in the area, equal to total building area / area of ​​the area; Bus stop density BSD: number of bus stops in a region / area of ​​the region; Road network density RD: total length of road network in the region / regional area; Industrial and commercial enterprise density BD: number of industrial and commercial enterprises in a region / regional area; The proportion of the tertiary industry in the region is TIP: the number of tertiary industry enterprises in the region / the total number of enterprises in the region; GDP per capita: GDP per capita in the region; The standardization of the observed variables is specifically: standardization is performed using the standard score Z-score, and the formula is as follows: Where x is the original data, μ and σ are the mean and standard deviation of the original data respectively; if the Z-score is positive, it means that the data point is higher than the mean; if the Z-score is negative, it means that the data point is lower than the mean; the larger the absolute value of the Z-score, the farther the data point is from the mean.

6. The method for evaluating driving factors of urban nighttime vitality according to claim 1, characterized in that: In step 3, the spatial structural equation model consists of a measurement model and a structural model.

7. The method for evaluating driving factors of urban nighttime vitality according to claim 6, characterized in that: The measurement model is used to verify whether the latent variables can be effectively represented by the corresponding observed variables. The formula is as follows: θ=λ θ f+e Where f represents the latent variable; θ represents the observed variable; λ θ represents the factor loading between the latent variable and the observed variable, expressing the correlation between the two; ε represents the measurement error; The structural model is used to reveal the causal paths between potential variables, analyze the correlation between potential variables, and quantify the impact of each factor on urban nighttime vitality. The calculation formula of the structural model is: f1=γ1f2+γ2f3+γ3f4+δ Among them, γ1, γ2, γ3 are unknown constants, f1, f2, f3, f4 represent latent variables, and it is assumed that δ has a univariate conditional autoregressive covariance structure with a covariance parameter of δ =(τ δ ,ρ δ ), where τ δ is the precision parameter, ρ δ is a spatially dependent parameter; The spatial factor model is introduced into the structural equation model, specifically: using the linear synergistic model to model (f2, f3, f4) T The joint distribution of: Among them, v1, v2, and v3 represent vectors of independent spatial processes in the study area. It is assumed that v1, v2, and v3 have a CAR covariance structure, and the overall scale parameter is set to 1. At the same time, the spatial correlation parameters are and It is a coefficient matrix that defines how the latent variables are generated by the spatial process. Each column represents the contribution of the spatial process to different latent variables. a1, a4, a7 represent the contribution to v1, a2, a5, a8 represent the contribution to v2, and a3, a6, a9 represent the contribution to v3.

8. The method for evaluating driving factors of urban nighttime vitality according to claim 1, characterized in that: In step 4, the spatial structural equation model verification and sensitivity analysis includes: model measurement, parameter analysis and model debugging.

9. The method for evaluating driving factors of urban nighttime vitality according to claim 8, characterized in that: The model calculation is specifically as follows: Using R language and Openbugs software, the calculated observed variables and neighborhood matrix were imported into the software for compilation and construction of a preliminary structural equation model. Based on the standardized path coefficients between the latent variables in Openbugs software, the influence of natural environment, social environment, and economic conditions on urban nighttime vitality was determined. The urban nighttime vitality index was also obtained based on the NTL and NTLD. The specific steps are as follows: using R language and Openbugs software to complete the model writing, importing the observed variables and spatial neighborhood matrix into Openbugs software as input data, and after the model is run, obtaining the standardized path coefficients between each potential variable and the urban nighttime vitality index. Based on the path coefficients, the influence of natural environment, social environment and economic conditions on urban nighttime vitality is analyzed, and an evaluation model of factors affecting urban nighttime vitality is constructed; Among them, the spatial neighborhood matrix is ​​specifically: the definition matrix like The matrix elements are defined as Otherwise ij =0, where n represents the number of grids, is the distance between grids i and j, θ is the distance threshold; The parameter analysis is specifically as follows: the evaluation indicators are calculated according to the spatial structural equation model, where the goodness of fit (GFI) value must be greater than 0.9, the factor loading of the latent variable and the observed variable must be greater than 0.7, and the residual spatial autocorrelation should be close to 0; The model debugging specifically includes: making necessary model corrections based on the results of the model evaluation; if the goodness of fit of the model does not meet the requirements, it is necessary to adjust the path coefficients and the relationship between the latent variables.

10. The method for evaluating driving factors of urban nighttime vitality according to claim 1, characterized in that: In step 5, the evaluation of the driving factors of urban nighttime vitality is specifically as follows: after the model evaluation and correction are completed, the urban nighttime vitality and various influencing factors are analyzed; based on the results of the spatial structural equation model, the impact of economic development, natural environment and social environment on urban nighttime vitality is quantified, and their spatial effects are analyzed; based on the analysis results, a scientific basis is provided for intervention policies and planning for urban nighttime vitality.

Images