Store operation management system based on big data analysis

Abstract

The present invention discloses a store operation management system based on big data analysis, which relates to the technical field of big data application and commercial operation management. The system includes: a multi-source data acquisition module that collects multi-source data inside and outside the store and marks anomalies; a data processing module that cleans data and extracts features such as store space and consumption trends; a regional trend modeling module that builds a model to quantify the store's sensitivity and response speed to regional trends and predict the impact of external events; a decision generation module that formulates adjustment plans and pushes them for execution; a model iteration module that collects data in real time, evaluates decision effects and provides feedback for optimization to form a closed loop; the present invention integrates multi-source data, extracts store features and external influences, uses a spatiotemporal model to quantify response speed, generates decision plans, and improves decision accuracy. At the same time, through real-time data acquisition and model iteration, a closed-loop management is formed to solve the problem of decision lag, so that stores can dynamically adapt to the market, predict and respond to external events in advance, and enhance competitiveness.

Description

Technical Field

[0001] The present invention relates to the technical field of big data application and commercial operation management, and specifically to a store operation management system based on big data analysis. Background Art

[0002] With the intensification of competition in the retail industry and the diversification of consumer demand, store operations management faces increasingly complex challenges. Traditional retail management methods have limitations in data integration, dynamic analysis and decision-making execution efficiency, making it difficult to effectively respond to the rapidly changing market environment.

[0003] At the data level, store operation-related data is usually scattered across different systems or platforms, lacking an efficient integration mechanism. At the same time, due to the lack of unified outlier identification rules, data quality is difficult to guarantee, which brings uncertainty to subsequent analysis. This data fragmentation and inconsistent quality problems restrict the comprehensive and accurate grasp of the store's operating status.

[0004] At the analytical level, existing technologies often rely on historical averages or static industry benchmarks. These methods struggle to dynamically quantify individual stores' differences in sensitivity to regional consumer trends, accurately capture the complex impact of external events on store operations and their time lags, and lack the ability to model dynamic transmission processes across time and space. This leads to discrepancies between forecasts and actual market responses, making it difficult to provide a basis for forward-looking decision-making.

[0005] At the decision-making and execution level, adjustment plans often rely on manual development and implementation, with low automation levels and insufficient integration with key business systems. This leads to delays between decision generation and implementation, making it difficult to allocate operational resources in a timely manner to respond to market changes, ultimately impacting sales conversion efficiency and customer experience.

[0006] The rapid development of big data and artificial intelligence technologies has provided new possibilities for the digital transformation of the retail industry. How to effectively integrate multi-source heterogeneous data, build a model that can dynamically simulate the transmission of consumption trends in the time and space dimensions, accurately quantify store response characteristics, and predict the impact of external events, and form a data-driven, closed-loop optimized intelligent decision-making execution mechanism, has become the key to improving the accuracy and adaptability of store operations management. Summary of the Invention

[0007] The purpose of the present invention is to make up for the shortcomings of the existing technology and provide a store operation management system based on big data analysis. It integrates sales, inventory, customer and external environment data through a multi-source data acquisition module, uses a rule engine to mark outliers to ensure data quality, and systematically constructs a spatiotemporal impact diffusion model to quantify the store's sensitivity and response speed to regional consumption trends. Combined with the prediction of the impact of external events, it generates product category adjustment, marketing strategy optimization and resource allocation plans. The decision generation module automatically connects to the ERP, SCM and CRM system execution plan, and continuously optimizes the model through an iterative feedback mechanism.

[0008] In order to solve the above technical problems, the present invention provides the following technical solutions: a store operation management system based on big data analysis, the system comprising:

[0009] Multi-source data collection module: This module collects data on store sales, inventory, customers, and basic attributes through smart POS terminals and IoT sensors. It also connects to the business district operation platform and official platforms to collect data on regional business district formats, customer flow, and community population. The data is then stored in a distributed data warehouse and abnormal data is marked.

[0010] Data processing and feature extraction module: extracts data from the data warehouse, cleans and standardizes it, extracts store space features, consumption trend features, and external event features, and outputs a structured data set;

[0011] Regional Trend Modeling Module: Based on structured data sets, this module constructs a spatiotemporal impact diffusion model within the geographic space of commercial districts to simulate the transmission of consumption trends across geographic space and time. By calculating each store's sensitivity index and response speed coefficient to regional trends, it quantifies the differences in store sensitivity and response speed to regional trends. Furthermore, through an external event impact prediction model, it deconstructs the influencing factors of external events, predicting their impact on store traffic, average customer spending, and category demand, as well as the time lag effect, and outputs the analysis results.

[0012] Decision generation and execution module: Based on the analysis results of the regional trend modeling module, it generates specific plans for product category adjustment, marketing strategy optimization, and operational resource allocation. It also pushes the decision plans to the store management terminal for display and automatically connects to the ERP, SCM, and CRM systems to trigger related operations.

[0013] Model iteration feedback module: collects store operation and regional dynamic data in real time, and uses online learning and offline retraining to iterate the model; collects operational data after decision execution every week, evaluates the decision effect from the sales, operation, and customer aspects, and optimizes the model based on the evaluation results to form an operational closed loop.

[0014] Furthermore, in the multi-source data acquisition module, the marking of abnormal data includes automatic marking through a preset rule engine, and the rule engine includes at least sales data abnormality rules, inventory data abnormality rules and external data abnormality rules. The sales data abnormality rules include situations where the amount of a single transaction exceeds the historical average by more than 3 times or the sales of a certain category drops by more than 50%; the inventory data abnormality rules include situations where the inventory quantity is negative, the shelf life is less than 10% and no warning is triggered, and the sensor collected data remains unchanged for a long time; the external data abnormality rules include situations where the passenger flow data of the business district suddenly becomes 0, the update time of the official planning document does not match the actual release time, and the traffic facility data is seriously inconsistent with the real-time road conditions.

[0015] Furthermore, in the data processing and feature extraction module, the specific contents of store space features, consumption trend features and external event features are as follows:

[0016] The store space features:

[0017] Geographic coordinate features: store latitude and longitude, straight-line distance to the business district center, number and location of competing products within 1 kilometer, and walking time to transportation hubs;

[0018] Spatial layout characteristics: store location, number of display shelves, entrance orientation;

[0019] The consumption trend characteristics are as follows:

[0020] Time series characteristics: daily / weekly / monthly growth rate of sales of each category, peak consumption period, and seasonal fluctuation coefficient;

[0021] Correlation features: product-related purchase patterns, correlation coefficients between average order value and customer traffic;

[0022] The characteristics of the external events are as follows:

[0023] Event type characteristics: the nature, scope of impact, and duration of the external event;

[0024] Event timeliness characteristics: the interval between the event occurrence time and the current time, the event warm-up period, and the event aftermath period.

[0025] Furthermore, in the regional trend modeling module, the spatiotemporal impact diffusion model is constructed, and its calculation formula is: ,in Represents a geographic location and time The intensity of consumption trends under It is the initial intensity coefficient of the trend, reflecting the intensity of the consumption trend at the starting point. It is set according to the peak value of historical trend data or industry experience. It is the current location The number of influential trend sources, It is The trend source position, It's location With trend source geographical distance, It is The effective influence radius of a trend source is obtained based on the statistical analysis of the range of trend propagation in historical data. It is the distance decay index, which is used to control the speed at which the trend strength decays with distance. The index value is obtained by curve fitting the historical trend diffusion data. It is the number of time-influencing factors, which is determined by analyzing the cyclical and seasonal characteristics of consumption trends over time. It is The weight of each time influencing factor is calculated by analyzing the influence of different time factors on consumption trends in historical data and using regression analysis. It is A time influence function is created, and the function form is set according to the characteristics of different time factors.

[0026] Furthermore, in the regional trend modeling module, the calculation formula for the sensitivity index is: ,in, It's a shop Sensitivity index to regional trends, It is the analysis time period, which is set according to research needs. It's a shop exist The rate of change of sales of a certain product category at a certain moment is calculated as follows: ,in and Stores exist Moment and The sales volume of the product category at the moment, the data is obtained from the store sales data collection part, All stores in the same area The average change rate of sales of this category at a certain moment in time is calculated by first calculating the change rate of each store and then finding the average. The formula is: ,in is the number of stores in the same area, yes The time weight of a moment is set according to the distance of the moment.

[0027] Furthermore, in the regional trend modeling module, the response speed coefficient is calculated using the following formula: ,in, It's a shop The response speed coefficient to regional trends, It's a shop The delay in sensing the change in regional trend is determined by comparing the time when the regional trend appears and the time when the store sales data begins to reflect the trend change. It's a shop Distance to the core source of the regional trend, It is the average effective distance of trend propagation. Through statistical analysis of the spread of historical trends in the region, the average propagation distance is calculated as a reference value.

[0028] Furthermore, in the regional trend modeling module, the external event impact prediction model is constructed, and its calculation formula is: ,in It is an external event About the store exist The impact value of passenger flow at each moment, It is an external event The number of sub-factors is determined by analyzing and breaking down the event. It is The influence of height on the store The influence weight of the store is calculated by principal component analysis using the influence of each sub-factor of similar events in historical data on the store. It is Factors affecting personality over time The change function is set according to the development stage of the event. It's a shop The distance to the place where the incident occurred is calculated by the geographical coordinates of the store and the place where the incident occurred. It's an event The spatial impact radius is determined by referring to the actual impact range of similar events and combining geographic information and market research data.

[0029] Furthermore, in the decision generation and execution module, the specific steps for generating specific plans for product category adjustment, marketing strategy optimization, and operational resource allocation based on the analysis results of the regional trend modeling module are as follows:

[0030] (1) Extracting data related to store sensitivity, response speed, and trend transmission from the regional trend modeling module;

[0031] (2) Based on the store characteristics and trend transmission results, formulate a product category adjustment plan and clarify the adjustment direction and proportion of products in different types of stores;

[0032] (3) Design marketing strategies based on the store’s response characteristics to trends;

[0033] (4) Based on the impact forecast of external events, plan the operational resource allocation plan for affected stores;

[0034] (5) Substitute each plan into the comprehensive decision weight formula and calculate the comprehensive weight of the plan based on the store-related data;

[0035] (6) By comparing the comprehensive weights of various options, the option with the highest weight is selected as the final implementation option;

[0036] (7) Submit the final plan to the management for review and adjust or confirm based on the feedback.

[0037] Furthermore, in the decision generation execution module, the calculation formula of the comprehensive decision weight formula is: ,in It is a decision-making plan About the store The comprehensive weight of They are sensitivity index, response speed coefficient, and weight adjustment parameters for the impact of customer flow changes. They are set through a multi-objective optimization algorithm based on store business objectives and regional market characteristics. It's a shop The sensitivity index, It's a shop The response speed coefficient, Is the external event affecting the store The impact of passenger flow, Decision-making options The scores in terms of sensitivity adaptation, response speed matching, and customer flow improvement are determined through analysis and evaluation of decision-making plans, combined with historical data analysis, and all candidate decision-making plans are traversed to select the store. The biggest plan.

[0038] Compared with existing technologies, this store operation management system based on big data analysis has the following beneficial effects:

[0039] 1. The present invention achieves comprehensive integration of sales, inventory, customer behavior and external environment data through a multi-source data acquisition module. Combined with data processing and feature extraction technology, it accurately depicts store spatial characteristics, consumption trends and the impact of external events. The regional trend modeling module uses a spatiotemporal influence diffusion model to quantify the store's sensitivity and response speed to regional consumption trends, generating scientific product category adjustment, marketing strategy optimization and resource allocation plans, thereby improving the accuracy of decision-making, enabling stores to quickly adapt to market changes, reducing trial and error costs, and achieving a dual improvement in operational efficiency and customer satisfaction, providing a data-driven intelligent management paradigm for the retail industry.

[0040] 2. The present invention continuously optimizes the prediction model and decision-making effects by collecting operational data in real time and adopting online learning and offline retraining methods. It conducts multi-dimensional evaluations of decision-making execution effects every week, forming a closed-loop management of "data collection-model prediction-decision execution-effect feedback". It solves the problems of delayed decision-making and passive adjustment in traditional retail management, enables store operations to have dynamic adaptability, and the system predicts the impact of external events on customer flow in advance and automatically generates plans to help stores seize market opportunities and ultimately achieve sustainable competitiveness improvement.

[0041] Other advantages, objects and features of the present invention will be described in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be learned from the practice of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort.

[0043] Figure 1 This is a flowchart of the store operation management system based on big data analysis;

[0044] Figure 2 This is a framework diagram of the store operation management system based on big data analysis. DETAILED DESCRIPTION

[0045] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the specific implementation methods, structures, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments.

[0046] Example 1:

[0047] Solutions to regional consumption trends in chain milk tea shops.

[0048] A milk tea chain brand has three stores in business district A in a second-tier city. A new subway line recently opened in the business district, and the occupancy rate of surrounding office buildings has increased. The brand hopes to optimize its store operation strategy through a system.

[0049] Multi-source data collection module: Collect milk tea sales data of each store for the past three months through smart POS terminals, including sales volume of each category and average customer spending; use IoT sensors to monitor inventory status, such as the remaining quantity and shelf life of pearls and milk caps; connect to the business district operation platform to obtain customer flow data of business district A, including average daily traffic volume and peak time distribution; collect officially released subway opening times and occupancy rate data of surrounding office buildings, store them in a distributed data warehouse, and mark abnormal data through a preset rule engine, such as the abnormally high single transaction amount of a store, the negative inventory quantity of a certain raw material, etc. Figure 1 shown.

[0050] Data processing and feature extraction module: Extracts collected data from the data warehouse, cleans and standardizes it, removes duplicate and erroneous data, and extracts store space features, such as the latitude and longitude of each store, the straight-line distance from the business district center, the number and location of competing milk tea shops within 1 km, the walking time to the subway entrance, and the floor where the store is located and the number of display shelves; extracts consumption trend features, including the week-on-week growth rate of milk tea sales of various categories, peak consumption periods (such as weekday afternoons and all day on weekends), seasonal fluctuation coefficients, and product-related purchase patterns (such as customers who buy pearl milk tea often pair it with pudding), and the correlation coefficient between average order value and customer flow; extracts external event features to clarify the nature of the external event of the subway opening (traffic benefit), the scope of impact (the entire business district A), the duration (long-term), the interval between the event occurrence time and the current time, the event warm-up period (one month before the subway opening), and the event aftermath period (expected three months after the opening).

[0051] Regional trend modeling module: Using the spatiotemporal impact diffusion model, the model is constructed based on the geographic space of business district A to simulate the transmission process of consumption trends in geographic space and time dimensions. The calculation formula of the spatiotemporal impact diffusion model is: ,in Represents a geographic location and time The intensity of consumption trends under It is the initial intensity coefficient of the trend, reflecting the intensity of the consumption trend at the starting point. It is set according to the peak value of historical trend data or industry experience. It is the current location The number of influential trend sources, It is The trend source position, It's location With trend source geographical distance, It is The effective influence radius of a trend source is obtained based on the statistical analysis of the range of trend propagation in historical data. It is the distance decay index, which is used to control the speed at which the trend strength decays with distance. The index value is obtained by curve fitting the historical trend diffusion data. It is the number of time-influencing factors, which is determined by analyzing the cyclical and seasonal characteristics of consumption trends over time. It is The weight of each time influencing factor is calculated by analyzing the influence of different time factors on consumption trends in historical data and using regression analysis. It is A time influence function is developed. The function form is set according to the characteristics of different time factors. The model is used to analyze the impact of subway opening and increased office building occupancy rates on consumption trends in business districts. The sensitivity index and response speed coefficient of each store to regional trends are calculated. The calculation formula for the sensitivity index is: ,in, It's a shop Sensitivity index to regional trends, It is the analysis time period, which is set according to research needs. It's a shop exist The rate of change of sales of a certain product category at a certain moment is calculated as follows: ,in and Stores exist Moment and The sales volume of the product category at the moment, the data is obtained from the store sales data collection part, All stores in the same area The average change rate of sales of this category at a certain moment in time is calculated by first calculating the change rate of each store and then finding the average. The formula is: ,in is the number of stores in the same area, yes The time weight of the moment is set according to the distance of time. The calculation formula of the response speed coefficient is: ,in, It's a shop The response speed coefficient to regional trends, It's a shop The delay in sensing the change in regional trend is determined by comparing the time when the regional trend appears and the time when the store sales data begins to reflect the trend change. It's a shop Distance to the core source of the regional trend, is the average effective distance of trend propagation. The average propagation distance is calculated as a reference value through statistical analysis of the spread of historical trends within a region. At the same time, the external event impact prediction model is used to break down the influencing factors of the subway opening and predict its impact on each store's customer flow, average customer spending, and category demand (for example, the demand for milk tea may increase during the afternoon tea period on weekdays), as well as the time lag effect. The calculation formula of the external event impact prediction model is: ,in It is an external event About the store exist The impact value of passenger flow at each moment, It is an external event The number of sub-factors is determined by analyzing and breaking down the event. It is The influence of height on the store The influence weight of the store is calculated by principal component analysis using the influence of each sub-factor of similar events in historical data on the store. It is Factors affecting personality over time The change function is set according to the development stage of the event. It's a shop The distance to the place where the incident occurred is calculated by the geographical coordinates of the store and the place where the incident occurred. It's an event The spatial impact radius of the event was determined by referring to the actual impact range of similar events, combining geographic information and market research data, and outputting analysis results. For example, the spatiotemporal impact diffusion model showed that the consumption trend intensity of Store A, close to the subway entrance, was expected to increase by 40%, while that of Store B, farther away, was expected to increase by 15%. The sensitivity index of Store A was 0.8 (higher than the regional average), and the response speed coefficient was 0.7 (short delay time). The external event impact prediction model indicated that the customer flow of Store A on weekday afternoons was expected to increase by 30%, while that of Store B was expected to increase by 10%. These values ​​clearly quantified the different impacts of the subway opening and the increase in office building occupancy rates on different stores, clarifying the direction for subsequent plan formulation.

[0052] Decision generation and execution module: Generate specific plans based on the modeling results of the regional consumption trend transmission effect. Extract the sensitivity, response speed, and trend transmission-related data of each store from the regional trend modeling module. Combine the characteristics of each store (for example, stores near subway entrances may have faster traffic growth) with the trend transmission results to formulate a product category adjustment plan, such as increasing the supply ratio of low-sugar and small-cup milk tea categories suitable for office workers. Design marketing strategies based on the response characteristics of each store to the trend, such as launching a "discount with subway tickets" campaign in the first week of the subway opening for stores near subway entrances to increase customer flow. Based on the impact of external events, plan operational resource allocation plans for affected stores, such as increasing staff shifts during peak hours and stocking up in advance. Substitute each plan into the comprehensive decision weight formula and calculate the comprehensive weight of the plan based on the store-related data. The calculation formula for the comprehensive decision weight formula is as follows: ,in It is a decision-making plan About the store The comprehensive weight of They are sensitivity index, response speed coefficient, and weight adjustment parameters for the impact of customer flow changes. They are set through a multi-objective optimization algorithm based on store business objectives and regional market characteristics. It's a shop The sensitivity index, It's a shop The response speed coefficient, Is the external event affecting the store The impact of passenger flow, Decision-making options The scores in terms of sensitivity adaptation, response speed matching, and customer flow improvement are determined through analysis and evaluation of decision-making plans, combined with historical data analysis, and all candidate decision-making plans are traversed to select the store. The largest plan is selected by comparing the comprehensive weights of various plans, and the plan with the highest weight is selected as the final implementation plan. It is submitted to the manager for review. After adjustment or confirmation based on feedback, the decision plan is pushed to the store management end for display, and automatically connected to the ERP, SCM, and CRM systems to trigger related operations, such as adjusting the procurement plan and updating the marketing activity settings. In view of the fact that the consumption trend intensity and customer flow of store A have increased significantly, and the peak is concentrated in the afternoon on weekdays, the supply ratio of "low-sugar small cup milk tea" is increased from 20% to 40%, and the supply of large cup equipment is reduced to 15%, which meets the needs of office workers; at the same time, taking advantage of its high sensitivity, it is opened in the subway In the first week, the company launched the "20% off with subway tickets" campaign and promoted corresponding product categories at subway entrances to quickly convert new customers. As the impact on store B was relatively weak, no major product category and marketing adjustments were made for the time being. Operational resource allocation needed to match the store's responsiveness with expected customer flow growth. Store A responded quickly and saw a significant increase in customer flow, so the company increased staff shifts from 3 to 5 pm on weekdays and prepared raw materials one hour in advance to ensure efficient operation. As store B had limited growth, it maintained its original shifts and inventory to avoid wasting resources. By matching numerical values ​​with actual operational needs, it achieved reasonable resource allocation and precise execution of strategies.

[0053] Model iteration feedback module: This module collects real-time operational and regional dynamic data from each store, such as actual customer flow, sales, and inventory consumption. It uses online learning and offline retraining to iterate the model. It collects operational data after decision execution every week and evaluates the effectiveness of decisions from the perspectives of sales (such as whether sales of each category have increased), operations (such as whether inventory turnover has increased), and customers (such as whether the proportion of new customers has increased). The module then optimizes the model based on the evaluation results to form an operational closed loop.

[0054] To sum up, the above-mentioned example of a chain milk tea shop shows that the store operation management system based on big data analysis can integrate internal and external information through multi-source data collection and feature extraction; with the help of the spatiotemporal impact diffusion model and the sensitivity index calculation model, it can quantitatively analyze the impact of regional trends; combined with the comprehensive decision weight formula to generate precise solutions, it realizes a closed loop from data collection to decision execution. The system can effectively respond to changes in business district traffic, help stores optimize product categories and marketing, improve sensitivity and response speed to regional trends, and form a data-driven operation optimization mechanism.

[0055] Example 2:

[0056] Response to external events and category optimization in fresh food supermarkets.

[0057] A fresh food supermarket is located near a large community. The community is about to hold a week-long cultural festival. At the same time, the weather has been unusually high recently. The supermarket needs to adjust its operating strategy to cope with possible changes in consumption.

[0058] Multi-source data collection module: Collect the supermarket's fresh food sales data for the past month through smart POS terminals, including sales volume and sales revenue of vegetables, fruits, and meat categories; use IoT sensors to monitor the temperature of cold storage and freezer storage, as well as the inventory quantity and shelf life of various fresh foods, such as Figure 2 As shown; connect to the business district operation platform to obtain the activity schedule and expected number of participants of the community cultural festival, collect high temperature weather warning information issued by the meteorological department, store it in the distributed data warehouse, and mark abnormal data through the preset rule engine. For example, the sales of a certain category of fresh food dropped by more than 50%, and the shelf life of a batch of fruit was less than 10% without triggering the warning.

[0059] Data processing and feature extraction module: Extract data from the data warehouse, clean and standardize it to ensure data accuracy, and extract store space features, such as the supermarket's latitude and longitude, the straight-line distance from the community center, the number and location of competing fresh supermarkets within 1 km, the walking time to the bus stop, and the floor where the supermarket is located, the number of display shelves, and the entrance direction; extract consumption trend characteristics, including the daily year-on-year growth rate of fresh food sales of each category, peak consumption periods (such as morning and evening every day), seasonal fluctuation coefficients, and commodity-related purchase patterns (such as customers who buy meat often pair it with vegetables), and the correlation coefficient between average customer spending and customer flow; extract external event features, clarify the nature of the external event of the community cultural festival (community activities), the scope of influence (residents in the surrounding communities), the duration (one week), the interval between the event occurrence time and the current time, the event warm-up period (three days before the cultural festival), and the event aftermath period (two days after the cultural festival). At the same time, clarify the nature of the external event of high temperature weather (weather factors), the scope of influence (the entire supermarket radiation area), the duration (estimated one week), and the interval between the event occurrence time and the current time.

[0060] Regional trend modeling module: Using the spatiotemporal impact diffusion model, a model is constructed based on the community where the supermarket is located and the surrounding geographical space to simulate the transmission process of community cultural festivals and high temperature weather on consumption trends. The calculation formula of the spatiotemporal impact diffusion model is: This model analyzes the potential increase in customer traffic brought about by community cultural festivals and the impact of hot weather on demand for fresh produce (for example, demand for heatstroke-fighting products like watermelon and mung beans may increase). The sensitivity index and response speed coefficient of supermarkets to regional trends are calculated using the following formula: , the calculation formula of the response speed coefficient is: Using the external event impact prediction model, we analyze the influencing factors of the community cultural festival and hot weather, and predict their impact and time lag effects on supermarket customer flow, average customer spending, and category demand (for example, sales of gift box fresh produce may increase during the cultural festival, and demand for cold drinks and cold dishes may increase in hot weather). The calculation formula of the external event impact prediction model is: , and output the analysis results.

[0061] Decision generation and execution module: Generate specific plans based on the modeling results of the regional consumption trend transmission effect. Extract the supermarket's sensitivity, response speed, and trend transmission-related data from the regional trend modeling module. Combined with the supermarket's characteristics (such as proximity to the community entrance and customer flow being greatly affected by the cultural festival) and the trend transmission results, formulate a product category adjustment plan, such as increasing the purchase volume of watermelon, mung beans, cold drinks, and cold dishes, and launching a fresh food combination packaged in cultural festival gift boxes. Design marketing strategies based on the supermarket's response to trends, such as holding a "Fresh Food Special Day" event during the cultural festival, offering discounts on heatstroke prevention and cooling products to attract customers. Plan operational resource allocation plans based on the impact of external events, such as increasing the number of cashiers and stock clerks during the cultural festival to ensure a smooth shopping experience. Substitute each plan into the comprehensive decision weight formula, and calculate the comprehensive weight of the plan based on the relevant supermarket data. The calculation formula for the comprehensive decision weight formula is as follows: , compare the comprehensive weights of each plan, select the plan with the highest weight as the final implementation plan, submit it to the manager for review, adjust or confirm it based on feedback, push the decision plan to the supermarket management end for display, and automatically connect to the ERP, SCM, and CRM systems to trigger related operations, such as adjusting purchase orders, updating price tags, and arranging employee shifts.

[0062] Model iteration feedback module: Real-time collection of supermarket operation and regional dynamic data, such as actual customer flow during the cultural festival, sales data of various categories of fresh produce, and inventory status. Using online learning and offline retraining iterative models, weekly collection of operational data after decision execution is carried out. The decision-making effect is evaluated from the aspects of sales (such as total sales, sales growth of various categories), operations (such as whether the inventory backlog rate has been reduced), and customers (such as whether customer satisfaction has improved). Feedback on the evaluation results is used to optimize the model to form an operational closed loop.

[0063] In summary, the application example of this fresh food supermarket shows that the system extracts store characteristics and consumption trends through multi-dimensional data collection and processing; uses the spatiotemporal impact diffusion model and external event impact prediction model to analyze the impact of cultural festivals and high temperature events on operations; generates adaptation plans through a comprehensive decision weight formula to achieve dynamic resource allocation. This process reflects the system's flexibility in responding to sudden external events, which can help supermarkets accurately capture changes in consumer demand, optimize category structure and marketing strategies, and continuously improve operational efficiency through model iteration, verifying the practicality and effectiveness of the system in fresh food retail scenarios.

[0064] The above description is merely a preferred embodiment of the present invention and does not constitute any form of limitation to the present invention. Although the present invention has been disclosed as above in terms of a preferred embodiment, it is not intended to limit the present invention. Any person skilled in the art can, without departing from the scope of the technical solution of the present invention, make some changes or modifications to equivalent embodiments using the technical contents disclosed above. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A store operation management system based on big data analysis, characterized by: The system includes: Multi-source data collection module: This module collects data on store sales, inventory, customers, and basic attributes through smart POS terminals and IoT sensors. It also connects to the business district operation platform and official platforms to collect data on regional business district formats, customer flow, and community population. The data is then stored in a distributed data warehouse and abnormal data is marked. Data processing and feature extraction module: extracts data from the data warehouse, cleans and standardizes it, extracts store space features, consumption trend features, and external event features, and outputs a structured data set; Regional trend modeling module: Based on the structured data set, the spatiotemporal impact diffusion model is constructed using the geographic space of the business district. The calculation formula of the spatiotemporal impact diffusion model is: ,in Represents a geographic location and time The intensity of consumption trends under It is the initial intensity coefficient of the trend, reflecting the intensity of the consumption trend at the starting point. It is set according to the peak value of historical trend data or industry experience. It is the current location The number of influential trend sources, It is The trend source position, It's location With trend source geographical distance, It is The effective influence radius of a trend source is obtained based on the statistical analysis of the range of trend propagation in historical data. It is the distance decay index, which is used to control the speed at which the trend strength decays with distance. The index value is obtained by curve fitting the historical trend diffusion data. It is the number of time-influencing factors, which is determined by analyzing the cyclical and seasonal characteristics of consumption trends over time. It is The weight of each time influencing factor is calculated by analyzing the influence of different time factors on consumption trends in historical data and using regression analysis. It is A time influence function is created, and the function form is set according to the characteristics of different time factors. The transmission process of consumption trends in geographical space and time dimensions is simulated. By calculating the store sensitivity index and response speed coefficient to regional trends, the differences in store sensitivity and response speed to regional trends are quantified. The external event impact prediction model is also used to decompose the influencing factors of external events and predict their impact on store customer flow, customer unit price, and category demand, as well as the time lag effect. The calculation formula of the external event impact prediction model is: ,in It is an external event About the store exist The impact value of passenger flow at each moment, It is an external event The number of sub-factors is determined by analyzing and breaking down the event. It is The influence of height on the store The influence weight of the store is calculated by principal component analysis using the influence of each sub-factor of similar events in historical data on the store. It is Factors affecting personality over time The change function is set according to the development stage of the event. It's a shop The distance to the place where the incident occurred is calculated by the geographical coordinates of the store and the place where the incident occurred. It's an event The spatial impact radius is determined by referring to the actual impact range of similar events, combining geographic information and market research data, and outputting analysis results; Decision generation and execution module: Based on the analysis results of the regional trend modeling module, it generates specific plans for product category adjustment, marketing strategy optimization, and operational resource allocation. It also pushes the decision plans to the store management terminal for display and automatically connects to the ERP, SCM, and CRM systems to trigger related operations. Model iteration feedback module: collects store operation and regional dynamic data in real time, and uses online learning and offline retraining to iterate the model; collects operational data after decision execution every week, evaluates the decision effect from the sales, operation, and customer aspects, and optimizes the model based on the evaluation results to form an operational closed loop.

2. The store operation management system based on big data analysis according to claim 1 is characterized in that: In the multi-source data acquisition module, the marking of abnormal data includes automatic marking through a preset rule engine, and the rule engine includes at least sales data abnormality rules, inventory data abnormality rules and external data abnormality rules. The sales data abnormality rules include situations where the amount of a single transaction exceeds the historical average by more than 3 times or the sales of a certain category drops by more than 50%; the inventory data abnormality rules include situations where the inventory quantity is negative, the shelf life is less than 10% and no warning is triggered, and the sensor collected data remains unchanged for a long time; the external data abnormality rules include situations where the passenger flow data of the business district suddenly becomes 0, the update time of the official planning document does not match the actual release time, and the traffic facility data is seriously inconsistent with the real-time road conditions.

3. The store operation management system based on big data analysis according to claim 1 is characterized in that: In the data processing and feature extraction module, the specific contents of store space features, consumption trend features, and external event features are as follows: The store space features: Geographic coordinate features: store latitude and longitude, straight-line distance to the business district center, number and location of competing products within 1 kilometer, and walking time to transportation hubs; Spatial layout features: The store's floor, number of display shelves, and entrance orientation; The consumption trend characteristics are as follows: Time series characteristics: daily / weekly / monthly growth rate of sales of each category, peak consumption period, and seasonal fluctuation coefficient; Correlation features: product-related purchase patterns, correlation coefficients between average order value and customer traffic; The characteristics of the external events are as follows: Event type characteristics: the nature, scope of impact, and duration of the external event; Event timeliness characteristics: the interval between the event occurrence time and the current time, the event warm-up period, and the event aftermath period.

4. The store operation management system based on big data analysis according to claim 1, characterized in that: In the regional trend modeling module, the calculation formula for the sensitivity index is: ,in, It's a shop Sensitivity index to regional trends, It is the analysis time period, which is set according to research needs. It's a shop exist The rate of change of sales of a certain product category at a certain moment is calculated as follows: ,in and Stores exist Moment and The sales volume of the product category at the moment, the data is obtained from the store sales data collection part, All stores in the same area The average change rate of sales of this category at a certain moment in time is calculated by first calculating the change rate of each store and then finding the average. The formula is: ,in is the number of stores in the same area, yes The time weight of a moment is set according to the distance of the moment.

5. The store operation management system based on big data analysis according to claim 1 is characterized in that: In the regional trend modeling module, the calculation formula for the response speed coefficient is: ,in, It's a shop The response speed coefficient to regional trends, It's a shop The delay in sensing the change in regional trend is determined by comparing the time when the regional trend appears and the time when the store sales data begins to reflect the trend change. It's a shop Distance to the core source of the regional trend, It is the average effective distance of trend propagation. Through statistical analysis of the spread of historical trends in the region, the average propagation distance is calculated as a reference value.

6. The store operation management system based on big data analysis according to claim 1, characterized in that: In the decision generation and execution module, the specific steps for generating specific plans for product category adjustment, marketing strategy optimization, and operational resource allocation based on the analysis results of the regional trend modeling module are as follows: (1) Extracting data related to store sensitivity, response speed, and trend transmission from the regional trend modeling module; (2) Based on store characteristics and trend transmission results, formulate a product category adjustment plan and clarify the adjustment direction and proportion of products in different types of stores; (3) Design marketing strategies based on the store’s response characteristics to trends; (4) Based on the impact forecast of external events, plan the operational resource allocation plan for affected stores; (5) Substitute each plan into the comprehensive decision weight formula and calculate the comprehensive weight of the plan based on the store-related data; (6) By comparing the comprehensive weights of various options, the option with the highest weight is selected as the final implementation option; (7) Submit the final plan to the management for review and adjust or confirm based on the feedback.

7. The store operation management system based on big data analysis according to claim 6 is characterized in that: In the decision generation execution module, the calculation formula of the comprehensive decision weight formula is: ,in It is a decision-making plan About the store The comprehensive weight of They are sensitivity index, response speed coefficient, and weight adjustment parameters for the impact of customer flow changes. They are set through a multi-objective optimization algorithm based on store business objectives and regional market characteristics. It's a shop The sensitivity index, It's a shop The response speed coefficient, Is the external event affecting the store The impact of passenger flow, Decision-making options The scores in terms of sensitivity adaptation, response speed matching, and customer flow improvement are determined through analysis and evaluation of decision-making plans, combined with historical data analysis, and all candidate decision-making plans are traversed to select the store. The biggest plan.

Images