Layout method of retail network space

Abstract

The invention provides a retail industry network space layout method, which comprises the following steps: on one hand, extracting consumer subjective initiative factors from multi-source space-time big data, and obtaining retail industry network space layout characteristics at the present stage from multi-source geographic big data; and a dynamic supply and demand relationship between different types of retail network space supplies and consumer space behavior demands in different time periods is obtained. On the other hand, a non-linear model is constructed by adopting a machine learning algorithm, and key influence factors and threshold values of the retail industry network space layout are determined. And finally, on the basis of the influence factors and the thresholds thereof, constructing a retail industry network space layout model through ArcGIS. Therefore, the retail network space layout method can improve the rationality and accuracy of the retail network space layout.

Description

Technical Field

[0001] The invention belongs to the technical field of retail network layout, and in particular relates to a layout method of retail network space. Background Art

[0002] The retail industry is an important part of the tertiary industry, playing a key role in promoting economic development, expanding domestic demand, and meeting the production and living needs of residents. With the development of science and technology, driven by the Internet, artificial intelligence and big data, the retail industry has also started digital transformation and upgrading, and the traditional physical retail consumption behavior model has gradually shifted to a new model that combines virtual online consumption with physical offline consumption behavior. As a retail commercial space carrier, retail outlets are not only a place for commodity transactions, but also an important platform for brands to communicate with customers, provide services and collect market information. However, there is currently a problem of mismatch, imbalance and even irrationality between residents' consumption behavior needs and retail outlet space supply, which has led to a decline in the economic benefits of retail outlets and a lack of economic vitality. This also shows that the development of retail outlets and their traditional layout model are no longer suitable for the development requirements of the digital age.

[0003] The applicant has found through research that the traditional spatial layout model of retail outlets has two main limitations, namely the limitations of the method for judging the rationality of the current spatial layout of retail outlets and the limitations of constructing a reasonable spatial layout model for retail outlets.

[0004] Specifically, on the one hand, the previous methods focusing on the rationality of the spatial layout of retail outlets mainly evaluated the degree of adaptation of the spatial layout of retail outlets to macro-static factors such as the level of economic development and urban construction pattern, and also mostly focused on the impact of policy planning, economy, market competition, location conditions, population, consumption capacity and other factors on the spatial layout of retail outlets, and based on this, proposed optimization strategies for the spatial layout of retail outlets. These factors mainly come from the official data of various statistical departments, and the statistical data disclosed by these government departments are lagging in time and can only be used to feedback information such as economic development, urban planning and construction, and population size at the objective static summary level. Static data lacks time dimension, cannot reflect changes, and may be outdated, which leads to the one-sidedness of the results of evaluating the rationality of the current spatial layout of retail outlets. On the other hand, the ordinary linear model of traditional econometric methods is currently mainly used to analyze the linear impact of these objective static factors on the spatial layout characteristics of retail outlets, which may lead to the inaccuracy of the construction of a reasonable spatial layout model for retail outlets. Summary of the invention

[0005] The purpose of the present invention is to solve the technical problem that in the above-mentioned prior art, only objective static data is used to analyze the rationality of the spatial layout of retail outlets, and the nonlinear influence of dynamic element information on the spatial layout of retail outlets is ignored, which leads to the unreasonable and inaccurate layout planning of the retail outlet space. A layout method for the retail outlet space is provided, which can improve the rationality and accuracy of the retail outlet space layout.

[0006] In order to solve the above problems, the present invention is implemented according to the following technical solutions:

[0007] The present invention provides a method for arranging a retail network space, the method comprising:

[0008] Cleaning and correcting the original data, including multi-source spatiotemporal big data, multi-source geographic big data and statistical data related to the spatial layout of retail outlets;

[0009] Extracting the consumer subjective initiative factors from the multi-source spatiotemporal big data after data cleaning and correction;

[0010] Based on the consumer subjective initiative factors, the dynamic supply and demand relationship between the space supply of different types of retail outlets and the consumer space behavior demand in different time periods is obtained;

[0011] Extracting the spatial layout characteristics of retail outlets from the multi-source geographic big data after data cleaning and correction;

[0012] Based on the statistical data after data cleaning and correction, objective static factors at the city level are extracted;

[0013] Building a database based on the subjective factors of consumers, the spatial layout characteristics of retail outlets and the objective static factors at the city level;

[0014] Using the data in the database as training samples to construct a nonlinear model;

[0015] By using the nonlinear model, key influencing factors and thresholds of the spatial layout of retail outlets are determined;

[0016] Based on the key influencing factors and their thresholds, a retail outlet spatial layout model is constructed using ArcGIS;

[0017] The layout of the retail outlet space is completed through the retail outlet space layout model.

[0018] Preferably, the multi-source spatiotemporal big data includes consumers' GPS trajectory data, mobile phone signaling data, social media data and APP data; the multi-source geographic big data includes point of interest data, high-resolution remote sensing data and street view image data; the statistical data related to the spatial layout of retail outlets includes statistical yearbooks, development bulletins and government public data.

[0019] Preferably, the method for correcting the original data includes: correcting the spatiotemporal characteristics related to residents' consumption behavior in the GPS trajectory data and mobile phone signaling data; correcting the geographical characteristics related to the spatial layout of retail outlets in the point of interest data; wherein the spatiotemporal characteristics include the consumer's departure spatial longitude and latitude, destination spatial longitude and latitude, start time, end time and duration; the geographical characteristics include longitude and latitude.

[0020] Preferably, a Python tool is called to extract the consumer subjective initiative factors in the multi-source spatiotemporal big data after data cleaning and correction; through GIS analysis or spatiotemporal visualization, the spatial layout characteristics of retail outlets in the multi-source geographic big data after data cleaning and correction are extracted; and a data statistical tool is called to extract objective static factors at the city level based on the statistical data after data cleaning and correction.

[0021] Preferably, the consumer subjective initiative factors include population mobility characteristics, spatiotemporal characteristics of residents' consumption behavior, consumption behavior needs and consumer selection preferences.

[0022] Preferably, the objective static factors at the city level include policy planning, economic development, urban construction, market competition, location conditions, land value, transportation conditions, population size and consumption capacity.

[0023] Preferably, the method of using the data in the database as training samples to construct the nonlinear model is: using a gradient boosting decision tree algorithm or a random forest algorithm to construct the nonlinear model.

[0024] Preferably, the method for determining the key influencing factors and thresholds of the spatial layout of retail outlets through a nonlinear model includes: after controlling other variables, analyzing the nonlinear impact of consumer subjective initiative factors and objective static factors at the city level on the spatial layout of retail outlets through the nonlinear model, and determining the key influencing factors and thresholds of the spatial layout of retail outlets; wherein the other variables include the consumer's gender, age group and employment status.

[0025] Preferably, the key influencing factors of the spatial layout of retail outlets include: consumers' subjective dynamic behavioral needs and the objective development of urban society.

[0026] Compared with the prior art, the present invention has the following beneficial effects:

[0027] The present invention provides a method for the layout of retail outlets space. On the one hand, by extracting consumer subjective initiative factors from multi-source spatiotemporal big data, the spatial layout characteristics of retail outlets at the current stage are obtained from multi-source geographic big data, thereby obtaining the dynamic supply and demand relationship between the spatial supply of different types of retail outlets and the consumer spatial behavior demand in different time periods. On the other hand, a machine learning algorithm is used to construct a nonlinear model to determine the key influencing factors and thresholds of the spatial layout of retail outlets. Finally, based on these influencing factors and their thresholds, a retail outlet spatial layout model is constructed through ArcGIS.

[0028] Therefore, the retail outlet space layout method of the present invention can improve the rationality and accuracy of the retail outlet space layout.

[0029] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The specific embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings, wherein:

[0031] Figure 1 It is a technical structure diagram of a layout method of a retail outlet space of the present invention;

[0032] Figure 2 It is an implementation flow chart of a layout method of a retail outlet space of the present invention. DETAILED DESCRIPTION

[0033] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0034] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0035] The terms used in this application are only for the purpose of describing specific embodiments and are not intended to limit this application. Unless otherwise defined, the technical terms or scientific terms used in this specification should be understood by people with ordinary skills in the field to which this application belongs. The words "first", "second" and similar words used in this specification and claims do not indicate any order, quantity or importance, but are only used to distinguish different technical features.

[0036] In order to solve the technical problem that the above-mentioned prior art only uses objective static data to analyze the rationality of the spatial layout of retail outlets, ignores the nonlinear influence of dynamic element information on the spatial layout of retail outlets, and thus leads to the unreasonable and inaccurate layout of the retail outlets space, a layout method of retail outlet space is provided, which can improve the rationality and accuracy of the spatial layout of retail outlets.

[0037] Figure 1 It is a technical structure diagram of a layout method of a retail outlet space of the present invention; Figure 2 It is an implementation flow chart of a layout method of a retail outlet space of the present invention.

[0038] As shown in the figure, a layout method of a retail outlet space according to the present invention includes:

[0039] Step 101: clean and correct the original data.

[0040] Specifically, the raw data include multi-source spatiotemporal big data, multi-source geographic big data, and statistical data related to the spatial layout of retail outlets.

[0041] It should be noted that multi-source spatiotemporal big data is related to multiple dimensions such as time, space, and attributes. It not only includes traditional geographic information data, but also covers new data generated by modern information technology such as the Internet of Things, mobile Internet, and social networks. These data are characterized by massive, multi-dimensional, and dynamically changing, and can be used to reveal the essential laws of things and conduct more comprehensive and accurate analysis and prediction. Therefore, the present invention uses multi-source spatiotemporal big data to participate in the analysis and construction of a retail outlet spatial layout model, which can obtain the dynamic supply and demand relationship between the retail outlet space supply and the consumer spatial behavior demand in different time periods, thereby clarifying the key element set that affects the retail outlet spatial layout, and then establishing a model to guide the reasonable layout of the retail outlet space.

[0042] In this embodiment, multi-source spatiotemporal big data includes consumers' GPS trajectory data, mobile phone signaling data, social media data (Weibo, Xiaohongshu, etc.) and APP data (Meituan, Dianping, etc.). Mobile phone signaling data determines the spatial location of users through information exchange between mobile phone users and base stations, and can relatively accurately record the spatiotemporal trajectory of population mobility. Mobile phone signaling data has the advantages of large sample size, low collection cost, and strong follow-up. It can be used to analyze residents' consumption characteristics. By designing a corresponding analysis framework, including data cleaning, trajectory point analysis, travel chain extraction, interest point analysis, and shopping method identification, users' consumption behaviors can be identified, such as consumption locations, consumption preferences, etc.

[0043] It should be noted that multi-source geographic big data focuses on the diversity and richness of geographic spatial data, including satellite remote sensing data, geographic information system (GIS) data, demographic data, traffic flow data, etc. These data sources can be used in urban planning, environmental monitoring, traffic management and other fields through fusion and analysis, helping decision makers to better understand the patterns and processes in geographic space. Therefore, the present invention can obtain the characteristics of the spatial layout of retail outlets at this stage by collecting and analyzing multi-source geographic big data, so as to better guide the construction of a spatial layout model of retail outlets.

[0044] In this embodiment, the multi-source geographic big data includes point of interest data, high-resolution remote sensing data, and street view image data. Point of interest data is data used in a geographic information system to identify a specific location or place, which may be considered "interesting" because of its specific purpose or significance. Points of interest can be any public place or private facility such as retail outlets, restaurants, theaters, parks, etc. Remote sensing data is data on the earth's surface acquired from a long distance by sensors on aircraft or satellites. High-resolution remote sensing data refers to remote sensing images with high spatial resolution, which can provide more detailed information on surface features. The resolution of this data can range from a few meters to a few centimeters, depending on the type of sensor and the flight altitude. Street view image data is a series of panoramic images acquired at street level, usually taken by a camera carried by a moving vehicle or on foot. These images provide an immersive visual experience that allows users to view the actual conditions of the streets and surrounding environment at different times and locations.

[0045] In this embodiment, the statistical data related to the spatial layout of retail outlets include statistical yearbooks, development bulletins, and government public data.

[0046] It should be noted that, in this embodiment, data cleaning of the raw data can ensure the data quality and accuracy of the analysis results. Correcting the raw data can improve the geolocation accuracy of the data and ensure that the data points are correctly mapped to the actual geographic location. On the other hand, accurate geographic data is the basis for effective spatial analysis, and data correction ensures the reliability of the analysis results. In addition, geographic data collected from different data sources or at different times may have inconsistent coordinate systems, and correction can unify these data into the same coordinate system. Therefore, data cleaning and correction are an indispensable part of the data science workflow, and they play a fundamental role in improving the quality of data analysis and the credibility of the results.

[0047] Specifically, data cleaning of raw data includes checking and cleaning the raw data, and excluding some data with missing information and abnormalities.

[0048] Specifically, the method for correcting the original data includes: (1) correcting the spatiotemporal characteristics of residents' consumption behavior in GPS trajectory data and mobile phone signaling data; (2) correcting the geographical characteristics of the spatial layout of retail outlets in the point of interest data; wherein the spatiotemporal characteristics include the consumer's departure spatial longitude and latitude, destination spatial longitude and latitude, start time, end time and duration; and the geographical characteristics include longitude and latitude.

[0049] Step 102: extract the consumer subjective initiative factors from the multi-source spatiotemporal big data after data cleaning and correction.

[0050] Specifically, Python tools are called to extract consumer subjective initiative factors from multi-source spatiotemporal big data after data cleaning and correction.

[0051] Specifically, factors of consumer initiative include population mobility characteristics, the temporal and spatial characteristics of residents' consumption behavior, consumption behavior needs and consumer choice preferences.

[0052] Step 103: Based on the subjective initiative of consumers, the dynamic supply-demand relationship between the space supply of different types of retail outlets and the space behavior demand of consumers in different time periods is obtained.

[0053] Step 104: extracting the spatial layout features of retail outlets from the multi-source geographic big data after data cleaning and correction.

[0054] Specifically, through GIS analysis or spatiotemporal visualization, the spatial layout characteristics of retail outlets in multi-source geographic big data after data cleaning and correction are extracted. Both GIS (Geographic Information System) analysis and spatiotemporal visualization can use geographic spatial data and temporal data to analyze and display spatial and temporal changes, helping decision makers and researchers better understand spatial and temporal phenomena and trends.

[0055] Step 105: Based on the statistical data after data cleaning and correction, objective static factors at the city level are extracted.

[0056] Specifically, call data statistical tools, based on the statistical data after data cleaning and correction, to extract objective static factors at the city level. Data statistical tools such as SPSS, Python, Power BI, etc.

[0057] Specifically, objective static factors at the city level include policy planning, economic development, urban construction, market competition, location conditions, land value, transportation conditions, population size and consumption capacity.

[0058] Step 106: Construct a database based on consumer subjective initiative factors, retail outlets spatial layout characteristics, and objective static factors at the city level.

[0059] Step 107: Use the data in the database as training samples to build a nonlinear model.

[0060] It should be noted that in the present invention, the purpose of constructing a nonlinear model for analysis is to capture and simulate complex, nonlinear relationships and patterns in the real world, which may involve multiple factors such as population distribution, customer behavior, geographical characteristics, and market competition. Due to its flexibility and strong predictive ability, the nonlinear model can more accurately reflect the interaction between these factors, thereby providing the retail industry with a more optimized network space layout strategy to ensure the maximization of service efficiency and customer satisfaction.

[0061] Specifically, a gradient boosting decision tree algorithm or a random forest algorithm is used to construct a nonlinear model.

[0062] It should be noted that the gradient boosted decision tree (GBDT) is an integrated learning algorithm that minimizes the loss function by gradually adding prediction weak learners (usually decision trees). The core idea of ​​GBDT is to optimize the loss function using the gradient descent method, and each iteration attempts to correct the residual of the previous iteration, thereby improving the prediction accuracy of the model. GBDT can handle various types of data, including numerical and categorical features. At the same time, because it is based on decision trees, the prediction process of the GBDT model is relatively easy to explain. Through parameters such as learning rate and tree depth, GBDT has a built-in regularization mechanism that helps prevent overfitting. Therefore, in the present invention, a nonlinear model is constructed by a gradient boosted decision tree algorithm, and a tree-based integrated method is used to construct multiple single decision trees, and then the results of these decision trees are combined. Through the tree-based integrated process, nonlinear effects can be explored and the relative importance of each factor in affecting the spatial layout characteristics of retail outlets can be quantified. After sorting them, they can be used to screen key factors with a greater degree of influence, and then clarify their accurate thresholds.

[0063] It should be noted that Random Forest (RF) is an integrated learning method that performs classification or regression prediction by constructing multiple decision trees. Since the random forest algorithm integrates multiple decision trees, the random forest can usually provide higher prediction accuracy. Prevent overfitting: Random forest reduces the risk of overfitting by increasing the diversity of the model. Random forest can be applied to classification and regression problems, and can process numerical and categorical data. Random forest can also evaluate the importance of each feature to the prediction result. Therefore, the present invention can also construct a nonlinear model through the random forest algorithm, analyze the nonlinear influence, and quantify the relative importance of each factor in affecting the spatial layout characteristics of retail outlets. After sorting them, they can be used to screen key factors with a greater degree of influence, thereby identifying their accurate thresholds.

[0064] Furthermore, in the random forest algorithm, the Bootstrap resampling method is used to generate a series of decision tree models {h(X, θ k ), k = 1, 2, 3, ...}, where X is the independent variable, θ is a random vector, and k represents the number of decision trees. This algorithm has the advantages of small prediction error, strong robustness, and effective prevention of variable collinearity. Assuming that the training set is an independent distribution set selected from random vectors Y and X, the mean square generalization error of any numerical prediction value h(X) is: E X，Y (Yh(X)) 2 The prediction value of random forest regression is {h(x, θ) of k trees. k )}, which has the following properties: When k→∞: EX，Y (Y-αv k h(X,θ k )) 2 →E X，Y (YE θ h(X,θ k )) 2 If for all θ, E(Y) = E X h(X,θ k ) then: Among them, PE * (tree)=E θ E X，Y (Yh(X,θ k )) 2 , Represents the residual Yh(X, θ k ) and Yh(X,θ′ k ), θ k and θ′ k Independent of each other.

[0065] It should be noted that the two important parameters of random forest, mtry and ntree, will directly affect the modeling effect. Among them, mtry specifies the number of variables selected at each branch of the decision tree. This parameter is generally selected as one-third of the number of variables in the regression model, which can effectively reduce the prediction error rate of the model. The parameter ntree refers to the number of decision trees contained in the random forest. This parameter is usually set to 500 or 1000 to achieve better modeling results.

[0066] Step 108: Determine the key influencing factors and their thresholds of the spatial layout of retail outlets through a nonlinear model.

[0067] Specifically, after controlling other variables, the nonlinear model is used to analyze the nonlinear impact of consumer subjective initiative factors and objective static factors at the city level on the spatial layout of retail outlets, and the key influencing factors and their thresholds of the spatial layout of retail outlets are determined; among them, other variables include consumers' gender, age group and employment status.

[0068] In this embodiment, the key influencing factors of the spatial layout of retail outlets include consumers' subjective dynamic behavioral needs and the objective development of urban society.

[0069] It is understandable that when data meets consumers' subjective dynamic behavioral needs and matches the objective development of urban society, the spatial layout of retail outlets can be better carried out.

[0070] Step 109: Based on the key influencing factors and their thresholds, a retail outlet spatial layout model is constructed using ArcGIS.

[0071] It should be noted that ArcGIS is a professional geographic information system (GIS) software developed by the Environmental Systems Research Institute (ESRI) of the United States, which is widely used in geographic data analysis, map making and spatial data management. In the present invention, a retail outlet spatial layout model that can target these key influencing factors and their thresholds is constructed through ArcGIS grid analysis. In ArcGIS, "grid" generally refers to a raster data format used for spatial analysis. Raster data is a data model in a geographic information system that divides the surface into small grids (pixels) of equal size, and each pixel stores one or more attribute values ​​(for example, elevation, vegetation coverage, temperature, etc.). The advantages of raster data include (1) continuity: raster data can represent continuously changing spatial phenomena, such as terrain, temperature distribution, etc.; (2) easy calculation: raster data is suitable for certain types of spatial analysis, such as pixel-based calculations, neighborhood analysis, etc.; (3) flexibility: raster data can easily perform spatial interpolation, classification and transformation operations; (4) visualization: the visualization effect of raster data is usually more intuitive and rich than that of vector data.

[0072] Step 110: Complete the layout of the retail outlet space through the retail outlet space layout model.

[0073] In summary, the present invention provides a method for the layout of retail outlets space. On the one hand, by extracting consumer subjective initiative factors from multi-source spatiotemporal big data, the spatial layout characteristics of retail outlets at the current stage are obtained from multi-source geographic big data, thereby obtaining the dynamic supply and demand relationship between the spatial supply of different types of retail outlets and the consumer spatial behavior demand in different time periods. On the other hand, a nonlinear model is constructed using a machine learning algorithm to determine the key influencing factors and thresholds of the spatial layout of retail outlets. Finally, based on these influencing factors and their thresholds, a spatial layout model of retail outlets is constructed using ArcGIS.

[0074] Therefore, the retail outlet space layout method of the present invention can improve the rationality and accuracy of the retail outlet space layout.

[0075] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0076] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for arranging retail outlet space, characterized in that: The method comprises: Cleaning and correcting the original data, including multi-source spatiotemporal big data, multi-source geographic big data and statistical data related to the spatial layout of retail outlets; Extracting the consumer subjective initiative factors from the multi-source spatiotemporal big data after data cleaning and correction; Based on the consumer subjective initiative factors, the dynamic supply and demand relationship between the space supply of different types of retail outlets and the consumer space behavior demand in different time periods is obtained; Extracting the spatial layout characteristics of retail outlets from the multi-source geographic big data after data cleaning and correction; Based on the statistical data after data cleaning and correction, objective static factors at the city level are extracted; Building a database based on the subjective factors of consumers, the spatial layout characteristics of retail outlets and the objective static factors at the city level; Using the data in the database as training samples to construct a nonlinear model; By using the nonlinear model, key influencing factors and thresholds of the spatial layout of retail outlets are determined; Based on the key influencing factors and their thresholds, a retail outlet spatial layout model is constructed using ArcGIS; The layout of the retail outlet space is completed through the retail outlet space layout model.

2. A method for arranging retail outlets space according to claim 1, characterized in that: The multi-source spatiotemporal big data includes consumers’ GPS trajectory data, mobile phone signaling data, social media data, and APP data; The multi-source geographic big data includes point of interest data, high-resolution remote sensing data and street view image data; The statistical data related to the spatial layout of retail outlets include statistical yearbooks, development bulletins and government public data.

3. A method for arranging retail outlets space according to claim 2, characterized in that: The method for performing data correction on original data comprises: Correcting the spatiotemporal characteristics of residents' consumption behaviors in the GPS trajectory data and mobile phone signaling data; Correcting the geographic features of the spatial layout of retail outlets in the POI data; Among them, the spatiotemporal characteristics include the consumer's departure spatial longitude and latitude, destination spatial longitude and latitude, start time, end time and duration; the geographical characteristics include longitude and latitude.

4. A method for arranging retail outlets according to claim 1, characterized in that: Calling Python tools to extract consumer subjective initiative factors from the multi-source spatiotemporal big data after data cleaning and correction; Extracting the spatial layout characteristics of retail outlets from the multi-source geographic big data after data cleaning and correction through GIS analysis or spatiotemporal visualization; The data statistics tool is called to extract objective static factors at the city level based on the statistical data after data cleaning and correction.

5. A method for arranging retail outlets space according to claim 4, characterized in that: The consumer subjective initiative factors include population mobility characteristics, temporal and spatial characteristics of residents' consumption behavior, consumption behavior needs and consumer choice preferences.

6. A method for arranging retail outlets space according to claim 4, characterized in that: The objective static factors at the city level include policy planning, economic development, urban construction, market competition, location conditions, land value, transportation conditions, population size and consumption capacity.

7. A method for arranging retail outlets space according to claim 1, characterized in that: The method of using the data in the database as training samples to construct a nonlinear model is: The nonlinear model is constructed using a gradient boosting decision tree algorithm or a random forest algorithm.

8. A method for arranging retail outlets space according to claim 1, characterized in that: The method for determining the key influencing factors and thresholds of the spatial layout of retail outlets through a nonlinear model includes: After controlling other variables, the nonlinear model is used to analyze the nonlinear impact of consumer subjective initiative factors and objective static factors at the city level on the spatial layout of retail outlets, and the key influencing factors and their thresholds of the spatial layout of retail outlets are determined; Among them, the other variables include the consumer's gender, age group and employment status.

9. A method for arranging retail outlets space according to claim 8, characterized in that: The key factors affecting the spatial layout of retail outlets include: Consumers' subjective dynamic behavioral demands and the objective development of urban society.

Images