Traffic venue screening methods, systems, and storage media for marketing purposes

Abstract

This application relates to the field of data processing technology, and discloses a method, system, and storage medium for screening marketing locations. The method includes: collecting multi-dimensional data through a monitoring system and calculating quality assessment indicators; extracting user characteristics to build profiles and calculating matching degrees; predicting effects by combining operational data; screening locations based on predictions and assessments; monitoring and analyzing the screening results, and dynamically adjusting to obtain an optimized solution. This application achieves multi-dimensional assessment of shopping mall quality, accurate matching of user profiles, and dynamic optimization of screening results, thereby improving the scientific nature and success rate of merchants' site selection decisions.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method, system and storage medium for filtering traffic venues for marketing. Background Technology

[0002] With the rapid development of the commercial real estate market, shopping malls, as important retail and consumption venues, have a crucial impact on merchants' business results through their site selection decisions. Traditional methods for selecting shopping mall locations mainly rely on manual experience and simple customer flow statistics. Merchants often only consider superficial factors such as basic customer flow and rent levels when choosing a location, lacking in-depth analysis of customer behavior characteristics and mall quality. While some existing data collection systems can record shopping mall customer flow data, and some shopping mall management software can provide basic customer flow analysis functions, most of these systems operate in isolation and fail to form a complete data analysis chain.

[0003] However, these existing technologies have significant shortcomings in application. First, data collection dimensions are limited, with most systems focusing only on foot traffic while neglecting customer quality. Second, the lack of precise user profiling makes it difficult to effectively assess the match between the mall's customer base and the needs of target merchants. Third, the absence of an objective site quality evaluation system leads to overly subjective site selection decisions. Finally, the selection results lack dynamic monitoring and optimization mechanisms, failing to adapt to changes in the market and foot traffic. These problems result in insufficient scientific rigor and accuracy in merchant site selection decisions, increasing operational risks. Summary of the Invention

[0004] This application provides a method, system, and storage medium for screening traffic venues for marketing, which enables multi-dimensional evaluation of shopping mall quality, accurate matching of user profiles, and dynamic optimization of screening results, thereby improving the scientific nature and success rate of merchants' site selection decisions.

[0005] Firstly, this application provides a method for screening traffic venues for marketing purposes. This method includes: collecting multi-dimensional data on the total number of visitors, unique visitors, dwell time, and bounce rate of a shopping mall's traffic venues through a deployed venue monitoring system; storing the data tagged according to source channel, device, and region to obtain an initial evaluation matrix; calculating a basic customer flow score, user stickiness coefficient, user engagement coefficient, and venue attribute coefficient based on the initial evaluation matrix; and obtaining a venue quality evaluation database through weighted calculation based on the quality evaluation database; and extracting user device type distribution, access time distribution, regional distribution, and consumption preference characteristics from the quality evaluation database to establish a venue user profile. The system calculates the matching score of the target merchant's user profile. Based on the matching score, combined with characteristics such as customer traffic, conversion rate, and spending amount in historical operating data, as well as site quality rating, business hours, and promotional activity formats, a site operation performance prediction score is obtained. Based on the operation performance prediction score, the site undergoes quality access assessment and user matching assessment. A grading standard is set according to the expected effect, and the site is screened in combination with the merchant site selection threshold to obtain the site screening combination result. Based on the site screening combination result, the site quality assessment indicators, user profile matching degree, and actual operating effect are monitored and analyzed. Dynamic adjustments are made through early warning triggering and effect comparison to obtain a shopping mall site optimization plan.

[0006] Secondly, this application provides a traffic venue screening system for marketing, the traffic venue screening system for marketing comprising:

[0007] The data collection module is used to collect multi-dimensional data on the total number of visitors, number of unique visitors, dwell time and bounce rate of shopping mall traffic areas by deploying a site monitoring system. The data is tagged and stored according to the source channel, device and region to obtain an initial evaluation matrix.

[0008] The calculation module is used to calculate the basic passenger flow score, user stickiness coefficient, user participation coefficient and site attribute coefficient based on the initial evaluation matrix, and obtain the site quality evaluation database through weighted calculation.

[0009] The extraction module is used to extract user device type distribution, access time distribution, geographical distribution and consumption preference characteristics based on the quality assessment database, establish site user profiles, and calculate the matching degree score of target merchant user profiles;

[0010] The prediction module is used to obtain a predicted score for the venue's operating performance based on the matching score, combined with features such as customer traffic, conversion rate, and consumption amount in historical operating data, as well as venue quality score, business hours, and promotional activities.

[0011] The evaluation module is used to conduct quality access assessment and user matching assessment of the venue based on the predicted operating performance score, set grading standards according to the expected results, and filter the venues by combining the merchant site selection threshold to obtain the venue selection combination results.

[0012] The monitoring module is used to monitor and analyze the quality assessment indicators, user profile matching degree and actual operating effect of the site based on the site selection and combination results, and make dynamic adjustments through early warning triggering and effect comparison to obtain the shopping mall site optimization plan.

[0013] A third aspect of the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described traffic site filtering method for marketing.

[0014] The technical solution provided in this application utilizes a site monitoring system to collect multi-dimensional data, achieving comprehensive monitoring of key indicators such as total mall visitors, number of unique visitors, dwell time, and bounce rate. This overcomes the limitations of traditional methods that only focus on single-dimensional customer flow. Furthermore, by employing tagging storage technology, attributes such as source channels, devices, and regions are structured and organized to form an information-rich initial evaluation matrix, laying a solid data foundation for subsequent analysis. Based on this evaluation matrix, the method introduces multi-dimensional evaluation indicators such as basic customer flow score, user stickiness coefficient, user engagement coefficient, and site attribute coefficient. Through weighted calculation, a site quality evaluation database is generated, achieving a scientific quantification of mall quality. Further, the distribution of user device types, access time periods, geographical distribution, and consumption preferences are extracted from the quality evaluation database to establish a three-dimensional site user profile. This profile is then used to calculate the user's profile against the target user profile using an intelligent matching algorithm. The matching score of merchant user profiles effectively solves the problem of traditional methods' difficulty in accurately assessing customer matching. Based on this, and combining features such as customer traffic, conversion rate, and spending in historical operating data, as well as factors such as site quality rating, operating hours, and promotional activities, a predictive model is applied to generate a predicted score for site operating performance, significantly improving the scientific rigor and accuracy of site selection decisions. Based on the predicted scores, the method designs a dual screening mechanism of quality access assessment and user matching assessment, and introduces expected effect grading standards and merchant site selection magnitude thresholds as constraints to obtain the optimal combination of site selection results. Finally, by establishing a monitoring and analysis closed loop, the quality assessment indicators, user profile matching degree, and actual operating performance of the selected sites are continuously monitored. Dynamic adjustments are made based on an early warning trigger mechanism and effect comparison analysis to generate mall site optimization plans, effectively addressing the challenges brought about by changes in the market environment and customer flow characteristics. The entire solution fully utilizes artificial intelligence algorithms for in-depth data mining and pattern recognition. In particular, advanced algorithms such as cluster analysis, similarity calculation, and regression prediction are introduced in the user profile construction, matching degree calculation, and business performance prediction stages. These algorithmic features play a key role in improving data processing accuracy, discovering potential correlation patterns, and achieving accurate prediction, significantly enhancing the efficiency and accuracy of shopping mall traffic site selection. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of one embodiment of a traffic venue screening method for marketing in this application.

[0017] Figure 2 This is a schematic diagram of one embodiment of a traffic venue screening system for marketing in this application. Detailed Implementation

[0018] This application provides a method, system, and storage medium for screening traffic locations in marketing. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0019] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the traffic venue screening method for marketing in this application includes:

[0020] Step S101: Collect multi-dimensional data on the total number of visitors, number of unique visitors, dwell time and bounce rate of the shopping mall traffic area by deploying a site monitoring system, and store the data in a tagged manner according to the source channel, device and region to obtain an initial evaluation matrix;

[0021] Step S102: Based on the initial evaluation matrix, calculate the basic passenger flow score, user stickiness coefficient, user participation coefficient, and site attribute coefficient, and obtain the site quality evaluation database through weighted calculation.

[0022] Step S103: Based on the quality assessment database, extract the distribution of user device type, access time period, geographical distribution and consumption preference characteristics, establish site user profiles, and calculate the matching score of the target merchant user profiles.

[0023] Step S104: Based on the matching score, combined with the characteristics of customer traffic, conversion rate, and consumption amount in historical operating data, as well as the venue quality score, business hours and promotional activities, the predicted score of venue operating effect is obtained.

[0024] Step S105: Based on the predicted operating performance score, conduct quality access assessment and user matching assessment of the site, set grading standards according to the expected results, and filter the sites in combination with the merchant site selection threshold to obtain the site selection combination results.

[0025] Step S106: Based on the site selection and combination results, monitor and analyze the site quality assessment indicators, user profile matching degree and actual operating effect, and make dynamic adjustments through early warning triggering and effect comparison to obtain the shopping mall site optimization plan.

[0026] It is understood that the executing entity of this application can be a traffic venue screening system used for marketing, or it can be a terminal or a server; the specific implementation is not limited here. This application's embodiment uses a server as an example for illustration.

[0027] Specifically, data is collected through the deployment of a site monitoring system. This system includes devices such as visitor counters, Wi-Fi probes, and facial recognition cameras, distributed at mall entrances, main corridors, and key areas on each floor. Visitor counters record the number of people entering and exiting, Wi-Fi probes capture mobile phone MAC addresses to identify unique visitors, and facial recognition cameras track customers' movement paths within the mall. Through these technologies, the system can accurately collect key indicators such as total visitor count, number of unique visitors, dwell time, and bounce rate. For example, a 50,000-square-meter mall with an average daily visitor flow of approximately 25,000 people, approximately 18,000 unique visitors, an average dwell time of 95 minutes, and a bounce rate (the percentage of visitors who only stay briefly in a single area before leaving) of 30%. The system tags and stores data based on the source channel (e.g., subway traffic, surrounding communities, online promotional activities), access device type (e.g., smartphones, wayfinding devices), and user location (based on IP address resolution or member registration information), forming an initial evaluation matrix.

[0028] Next, several key indicators were calculated based on the initial evaluation matrix. The basic visitor flow score is a weighted combination of the total number of visitors and the number of unique visitors after standardization, reflecting the basic traffic level of the venue. The user stickiness coefficient is calculated by comparing the dwell time with the industry standard value, representing customers' willingness to stay at the venue. The user engagement coefficient is derived from the bounce rate data, reflecting the level of customer activity at the venue. The venue attribute coefficient integrates data from three dimensions: source channel credibility, equipment coverage, and geographical distribution. These four indicators, after weighted calculation, form a venue quality assessment database, providing a foundation for subsequent analysis.

[0029] Based on the quality assessment database, the system analyzes and extracts user characteristics. Device type distribution reflects the proportion of consumers using mobile devices, wayfinding equipment, and membership cards; visit time distribution shows the temporal patterns of peak and off-peak traffic; geographic distribution indicates the geographical breadth of customer origins; and consumption preferences are determined by analyzing purchase records to identify customer preferences for specific goods or services. These characteristics collectively constitute the venue's user profile. Simultaneously, target merchants provide standard user profiles based on their own customer characteristics. The system quantifies the matching score by calculating the Euclidean distance between the two profiles across various dimensions, judging the degree of fit between the venue's customer base and the merchant's target audience. The matching score is combined with historical operating data for performance prediction. Historical operating data includes indicators such as foot traffic, conversion rate (the proportion of actual consumers to total foot traffic), and average transaction value. The system also considers factors such as venue quality rating, operating time characteristics, and promotional activity formats (e.g., discount type, intensity, and scope) to build a predictive model and derive a predicted venue performance score, used to assess the potential performance of merchants after they move into the venue.

[0030] Based on the predicted scores, the system performs quality access assessment and user matching assessment. The quality access assessment sets a minimum entry threshold to screen out venues that meet the basic requirements; the user matching assessment ensures that the customer base and merchant needs are highly aligned; the expected effect grading divides venues into high, medium, and low levels; and the deployment volume threshold sets the site selection range based on the merchant's budget. After cross-validation of these standards, a combination of venue selection results is generated.

[0031] The system continuously monitors and dynamically optimizes site performance. Monitoring includes fluctuations in quality assessment indicators, changes in user profiles, and discrepancies between actual and expected operating results. When indicators exceed warning thresholds, the system automatically triggers an alert mechanism. Through performance comparison and dynamic adjustments, the optimization plan is continuously improved. This entire process forms a closed loop, ensuring that the mall traffic site selection results always align with merchant needs and market changes. For example, when a clothing brand is looking for a new location, the system analyzes data from 50 malls in the city. Initial evaluation matrix analysis identifies 30 malls with sufficient basic customer traffic. Combining the brand's target customer characteristics (18-35 years old, middle-to-high income, preference for fashion items) with the user profiles of each mall, the matching degree is calculated, further narrowing the scope to 15 malls. After analyzing historical operating data, the system predicts the 5 most promising sites and continuously monitors the performance indicators of these malls to ensure the timeliness and accuracy of the recommendations.

[0032] In this embodiment, a site monitoring system is deployed to collect multi-dimensional data, enabling comprehensive monitoring of key indicators such as total mall visitors, number of unique visitors, dwell time, and bounce rate. This overcomes the limitations of traditional methods that only focus on single customer flow. Furthermore, tagging storage technology is used to structure attributes such as source channels, devices, and regions, forming an information-rich initial evaluation matrix, laying a solid data foundation for subsequent analysis. Based on this evaluation matrix, the method introduces multi-dimensional evaluation indicators such as basic customer flow score, user stickiness coefficient, user engagement coefficient, and site attribute coefficient. A site quality evaluation database is generated through weighted calculation, achieving a scientific quantification of mall quality. Further, the distribution of user device types, access time periods, geographical distribution, and consumption preferences are extracted from the quality evaluation database to establish a three-dimensional site user profile. An intelligent matching algorithm is then used to calculate the matching with target merchants. The matching score of user profiles effectively solves the problem of traditional methods' difficulty in accurately assessing customer matching. Based on this, and combining features such as customer traffic, conversion rate, and spending in historical operating data, as well as factors such as site quality rating, operating hours, and promotional activities, a predictive model is applied to generate a predicted score for site operating performance, significantly improving the scientific rigor and accuracy of site selection decisions. Based on the predicted scores, the method designs a dual screening mechanism of quality access assessment and user matching assessment, and introduces expected performance grading standards and merchant site selection thresholds as constraints to obtain the optimal combination of site selection results. Finally, by establishing a closed-loop monitoring and analysis system, the quality assessment indicators, user profile matching scores, and actual operating performance of the selected sites are continuously monitored. Dynamic adjustments are made based on an early warning trigger mechanism and performance comparison analysis to generate optimized shopping mall site plans, effectively addressing the challenges brought about by changes in the market environment and customer flow characteristics. The entire solution fully utilizes artificial intelligence algorithms for in-depth data mining and pattern recognition. In particular, advanced algorithms such as cluster analysis, similarity calculation, and regression prediction are introduced in the user profile construction, matching degree calculation, and business performance prediction stages. These algorithmic features play a key role in improving data processing accuracy, discovering potential correlation patterns, and achieving accurate prediction, significantly enhancing the efficiency and accuracy of shopping mall traffic site selection.

[0033] In one specific embodiment, the process of performing step S101 may specifically include the following steps:

[0034] (1) Collect the passenger flow records of each shopping mall traffic area within a specified time window through the monitoring system to obtain the total number of visitors to the shopping mall traffic area;

[0035] (2) Deduplicate the user identification information in the total number of visitors to the shopping mall to obtain the number of unique visitors to the shopping mall.

[0036] (3) By recording the entry and exit times of users in the shopping mall traffic area, the time difference of each visit is calculated to obtain the dwell time in the shopping mall traffic area;

[0037] (4) Calculate the ratio of the number of users who only access a single area in the shopping mall traffic area to the total number of users who access the area, and obtain the shopping mall traffic area bounce rate.

[0038] (5) Extract source channel identifiers based on user access information and classify them according to online traffic, surrounding communities, business cooperation and brand activities to obtain source channel data for shopping mall traffic venues;

[0039] (6) By parsing the user's device identification information, the data on the access devices of the shopping mall traffic area is obtained by classifying and statistically analyzing the mobile terminal, shopping mall wayfinding equipment and membership card.

[0040] (7) Perform geographic location analysis based on user location information, and conduct distribution statistics according to province, city and region to obtain user geographic data of shopping mall traffic venues;

[0041] (8) Link the source channel data, access device data and user geographic data of shopping mall traffic venues in multiple dimensions, establish a tag index system, and obtain an initial evaluation matrix.

[0042] Specifically, a monitoring system collects customer flow records within a specified time window in the shopping mall's traffic areas to obtain total visitor data. This monitoring system refers to a collection of hardware devices, including customer flow counting devices, wireless probes, and cameras, deployed at various entrances, main passages, and areas of the mall. The specified time window is typically a daily operating period, such as 10:00 AM to 10:00 PM, but can also be set to other time dimensions such as weeks or months as needed. Customer flow records refer to raw data such as the time customers enter the mall and their movement trajectories. Total visitors refer to the total number of people entering the mall within the time window, including repeat entries. The user identification information in the total visitor count is deduplicated to obtain the number of unique visitors. User identification information includes, but is not limited to, mobile phone MAC addresses captured by Wi-Fi probes, Bluetooth device IDs, membership card numbers, and temporary IDs generated by the facial recognition system. Deduplication uses algorithms such as hash table mapping to merge multiple visits by the same user into a single record, thereby calculating the number of different customers actually visiting within the specified time window, i.e., the number of unique visitors. By recording the entry and exit times of users in the shopping mall's traffic areas, the time difference of each visit is calculated to obtain the dwell time. Entry time refers to the timestamp when a customer is first captured by the monitoring system, and exit time refers to the timestamp when a customer is last captured by the system. The time difference is calculated by subtracting the timestamps, typically in minutes. For customers who enter and exit the mall multiple times, each entry and exit is calculated separately, and the average is calculated. Dwell time reflects a customer's willingness to stay in the mall and their shopping experience, and is an important indicator for assessing the mall's attractiveness. The bounce rate is calculated by dividing the number of users who only visited a single area within the mall's traffic area by the total number of users. A single area refers to a functional zone within the mall, such as the lobby, elevator entrance, or a specific counter. Visiting only a single area means that a customer enters the mall, stays briefly in one area, and leaves without exploring other areas. The bounce rate is calculated by dividing the number of users who visited a single area by the total number of users, reflecting the customer's responsiveness to the mall's overall attractiveness; a higher rate indicates a weaker ability to attract and retain customers.

[0043] Source channel identifiers are extracted based on user visit information and categorized into online traffic, surrounding community, business partnerships, and brand activities to obtain source channel data. Source channel identifiers can be obtained through various means: the source page from scanning the mall's QR code, the distribution channel for coupons, and channel markers during member registration. Online traffic refers to channels that guide customers to the store through online platforms, such as social media promotion and search engine advertising; surrounding community refers to natural foot traffic from residential and office areas around the mall; business partnerships refer to foot traffic driven by joint marketing with other businesses; and brand activities refer to foot traffic attracted by promotional activities, exhibitions, performances, etc., held by the mall. Source channel data is crucial for understanding customer acquisition methods and evaluating marketing effectiveness.

[0044] By analyzing user device identification information and categorizing it into mobile terminals, mall wayfinding devices, and membership cards, access device data is obtained. Device identification information includes attributes such as device type, brand, and model. Mobile terminals refer to the smartphones used by customers, which can be captured via Wi-Fi probes or Bluetooth signals; mall wayfinding devices refer to facilities such as interactive screens and information kiosks within the mall; membership cards include physical membership cards and electronic membership cards. Access device data reflects the ways customers interact with the mall and their technological preferences, providing guidance for optimizing service facilities and improving user experience. Geographic location data is obtained by analyzing user location information and distributing it by province, city, and region. Location information can be obtained from sources such as IP addresses, mobile GPS signals, and membership registration information. Geographic location analysis uses technologies such as GeoIP to map location information to an administrative division system. User regional data can show the geographical distribution of the mall's customer base, which is of great value in understanding the mall's reach and influence.

[0045] By linking data on shopping mall traffic sources, access devices, and user geographic locations across multiple dimensions, a tag index system is established to obtain an initial evaluation matrix. Multi-dimensional linking refers to associating these data types according to user IDs to form a complete user access profile. The tag index system is a structured data organization method that categorizes and indexes user attributes and behavioral characteristics in the form of tags. The initial evaluation matrix is ​​a multi-dimensional data table where rows represent different users or user groups, columns represent various indicator characteristics, and the values ​​in the matrix represent the performance or attributes of a specific user on specific characteristics.

[0046] In one specific embodiment, the process of performing step S102 may specifically include the following steps:

[0047] (1) Normalize the total number of visitors and the number of unique visitors in the initial evaluation matrix, and then combine them linearly in a ratio of 6:4 to obtain the basic visitor flow score.

[0048] (2) Convert the dwell time data in the initial evaluation matrix into minutes and calculate the user stickiness coefficient by comparing it with the industry average dwell time.

[0049] (3) The bounce rate data in the initial evaluation matrix is ​​reverse-converted, and the user engagement coefficient is obtained by calculating the degree of deviation from the industry benchmark bounce rate.

[0050] (4) Based on the source channel data in the initial evaluation matrix, weighted credibility scores are given to each channel to obtain the source channel credibility;

[0051] (5) Based on the access device data in the initial evaluation matrix, the number and distribution ratio of the covered device types are statistically analyzed to obtain the device coverage range;

[0052] (6) Based on the user geographic data in the initial evaluation matrix, calculate the concentration and coverage of geographic distribution to obtain the geographic distribution;

[0053] (7) Linearly combine the credibility of the source channel, the coverage of the equipment, and the geographical distribution to obtain the site attribute coefficient;

[0054] (8) The basic passenger flow score, user stickiness coefficient, user participation coefficient and site attribute coefficient are weighted and summed to obtain the site quality assessment database.

[0055] Specifically, the total visitor data and unique visitor data in the initial evaluation matrix are normalized to eliminate differences between data of different magnitudes. The normalization process uses a maximum-minimum standardization method, mapping the original data to the [0,1] interval. Specifically, for the total visitor data, the maximum and minimum values ​​from the mall's historical data or industry standards are used as reference points to calculate the relative position of the current total visitor count within this range; the same method is applied to the unique visitor data. After normalization, both are linearly combined according to a 6:4 weighting ratio to form the basic customer flow score. This 6:4 ratio reflects the higher importance of unique visitors relative to total visitors in the mall evaluation, as unique visitors better reflect the actual customer base size covered by the mall.

[0056] Converting dwell time data in the initial evaluation matrix to minutes is to standardize the measurement and facilitate calculation and comparison. The original dwell time data may have been recorded in seconds or hours, requiring a unified conversion to minutes. The user stickiness coefficient is calculated by comparing it to the industry average dwell time. The industry average dwell time is a benchmark value determined based on factors such as mall type, size, and positioning. The user stickiness coefficient reflects customers' willingness to stay in the mall compared to industry standards; a higher coefficient indicates stronger mall attractiveness and customer retention capabilities.

[0057] The bounce rate data in the initial evaluation matrix is ​​inversely transformed because the bounce rate is a negative indicator; a higher bounce rate indicates lower venue attractiveness. The purpose of the inverse transformation is to convert it into a positive indicator, facilitating weighted calculations with other indicators. The specific calculation formula is as follows:

[0058]

[0059] Where PEC represents the Participation Engagement Coefficient, BR represents the current bounce rate of the shopping mall, and BR base BR represents the industry benchmark bounce rate. max The theoretical maximum bounce rate (usually set to 100%) is represented by λ, which is an adjustment coefficient used to control the sensitivity of switching. When the mall's bounce rate equals the industry benchmark, the PEC value is 1; when the mall's bounce rate is higher than the industry benchmark, the PEC value is less than 1; and when the mall's bounce rate is lower than the industry benchmark, the PEC value is greater than 1. This calculation method ensures that the user engagement coefficient accurately reflects the mall's relative performance in retaining customers.

[0060] Based on the source channel data in the initial evaluation matrix, a weighted credibility score is assigned to each channel. Different source channels have different values; for example, customers attracted through brand events typically have higher brand loyalty and willingness to spend, while natural foot traffic from surrounding communities is more stable. The weighted credibility score is determined by weighting coefficients based on factors such as stability, conversion rate, and customer value for each channel. By multiplying the foot traffic share of each channel by its weighting coefficient and summing the results, the source channel credibility index is obtained, reflecting the overall quality and stability of the mall's foot traffic sources.

[0061] Based on the access device data in the initial evaluation matrix, the number and distribution ratio of covered device types are statistically analyzed to obtain the device coverage range. The number of device types reflects the diversity of touchpoints between the mall and customers, while the distribution ratio reflects the balance among various touchpoints. Device coverage range is an indicator for measuring the mall's omnichannel service capabilities. The calculation considers both the number of supported device types and the balance of usage among different devices, typically quantified using information entropy or similar diversity indicators.

[0062] Based on the user geographic data in the initial evaluation matrix, the concentration and coverage of geographic distribution are calculated to obtain geographic distribution indicators. Concentration reflects the degree of customer aggregation and is usually calculated using the Gini coefficient or similar inequality measures; coverage reflects the geographical scope of the mall's influence and is usually calculated by counting the number of administrative divisions covered by the effective customer base and combining distance factors. The geographic distribution indicators comprehensively consider these two aspects, focusing on both the mall's deep penetration in the core business district and its radiation capacity in a wider area.

[0063] The site attribute coefficient is obtained by linearly combining the credibility of the source channel, the coverage of equipment, and the geographical distribution. Linear combination refers to assigning weights to the three indicators and then calculating a weighted average. The weighting is typically determined based on the shopping mall's strategic priorities and operational objectives. For example, high-end shopping malls that prioritize customer quality might assign a higher weight to the credibility of the source channel, while regional shopping malls seeking broad market coverage might place more emphasis on geographical distribution indicators. The site attribute coefficient reflects the shopping mall's overall performance in terms of customer structure.

[0064] The site quality assessment database is obtained by weighted summation of basic passenger flow scores, user stickiness coefficients, user participation coefficients, and site attribute coefficients. Weighted summation is a commonly used aggregation method in multi-criteria decision-making, which can synthesize the performance of various dimensions to form an overall evaluation. The weight allocation is usually determined based on expert experience and historical data analysis, or it can be derived through scientific methods such as the analytic hierarchy process (AHP). The site quality assessment database not only contains the final comprehensive score, but also retains the original data and intermediate calculation results of each dimension, facilitating subsequent analysis and application.

[0065] Taking the evaluation of a shopping mall as an example, the mall has an average daily foot traffic of 15,000 people and 9,000 unique visitors. Dividing these two figures by the maximum values ​​of similar malls in the area (20,000 and 12,000 people respectively) yields normalized values ​​of 0.75 and 0.75. Combining these values ​​in a 6:4 ratio, the basic foot traffic score is 0.75 × 0.6 + 0.75 × 0.4 = 0.75. The average dwell time of customers in this shopping mall is 105 minutes, which is 1.17 higher than the industry average of 90 minutes, indicating a user stickiness coefficient of 1.17. The mall's bounce rate is 25%, the industry benchmark is 35%, the theoretical maximum is 100%, and the adjustment coefficient λ is taken as 0.8. Substituting these values ​​into the formula, the user engagement coefficient is calculated as 1 + (0.35 - 0.25) / (1 - 0.35) × 0.8 = 1.12. Among the source channels, brand activities account for 30%, surrounding communities for 40%, online traffic for 20%, and business partnerships for 10%, with weighting coefficients of 1.2, 1.0, 0.9, and 1.1 respectively. The weighted average yields a source channel credibility of 1.05. The shopping center supports three device types: mobile terminals, mall wayfinding devices, and membership cards. The usage ratio is relatively balanced, resulting in a device coverage index of 0.95. The customer source area is concentrated in three major administrative regions, with moderate coverage, resulting in a geographical distribution index of 0.85. A linear combination of these three indicators with a weighting ratio of 3:2:5 yields a site attribute coefficient of 0.945. Finally, the indicators of the four dimensions were weighted and summed according to their respective importance. Assuming the weights were 0.35, 0.25, 0.25 and 0.15 respectively, the comprehensive score was obtained as 0.75×0.35+1.17×0.25+1.12×0.25+0.945×0.15=0.99625. This score and related detailed data were stored in the site quality assessment database.

[0066] In one specific embodiment, the process of executing step S103 may specifically include the following steps:

[0067] (1) Extract the device type data of users in the shopping mall traffic area from the quality assessment database, and perform statistical calculations on the distribution ratio of mobile devices, wayfinding terminals and membership cards to obtain the device type distribution;

[0068] (2) Extract user access time data from the quality assessment database, divide the time intervals into 24-hour intervals, and count the access frequency of each time interval to obtain the access time distribution.

[0069] (3) Extract user location data from the quality assessment database, calculate the proportion of users accessing the provincial and municipal regions, and obtain the geographical distribution;

[0070] (4) Extract user consumption data from the quality assessment database, and perform weighted calculations on consumption categories, amount levels and purchase frequency to obtain consumption preferences;

[0071] (5) Combine the distribution of device type, access time, geographical distribution and consumption preferences in multiple dimensions, and establish weight coefficients to obtain the site user profile;

[0072] (6) Construct the target merchant customer group characteristics by demographic characteristics, consumption capacity level and category preference dimension to obtain the target merchant user profile;

[0073] (7) Calculate the feature vectors of the corresponding dimensions of the site user profile and the target merchant user profile, and calculate the dimension difference value through the Euclidean distance formula to obtain the matching score.

[0074] Specifically, establishing user profiles and calculating matching degrees are key steps in the mall traffic area screening method. Extracting device type data from the quality assessment database refers to classifying and statistically analyzing the terminal devices used by users during the access data analysis process. Device type data includes three main categories: mobile devices (smartphones, tablets, etc.), wayfinding terminals (self-service navigation devices, interactive screens, etc. within the mall), and membership cards (physical or electronic membership cards). Statistically calculating the distribution percentage of these devices requires first counting the number of times each type of device was used, then dividing by the total number of visits to obtain the percentage value. For example, data captured by Wi-Fi probes and Bluetooth signals shows that out of 10,000 visits to a certain mall, mobile devices accounted for 8,500 visits, wayfinding terminals for 1,000 visits, and membership cards for 500 visits, resulting in distribution percentages of 85%, 10%, and 5%, respectively. These values ​​constitute the device type distribution, reflecting the main ways users interact with the mall and their technological preferences. Extracting user access time data from the quality assessment database is for analyzing the time patterns of customer visits to the mall. The mall is divided into 24-hour time intervals, typically in one-hour increments, i.e., 0-1 AM, 1-2 AM... 11 PM-12 AM, for a total of 24 intervals. The frequency of visits within each time interval is statistically analyzed, calculating how many customers enter the mall during each time period. This data can be obtained through entrance flow counting devices, camera recognition systems, or Wi-Fi probes. Dividing the number of visitors in each time period by the total number of visitors yields the visit percentage for each time period, thus forming the visit distribution by time period. This distribution clearly shows the mall's peak and off-peak traffic periods, providing a reference for merchants in site selection and business hour scheduling.

[0075] Extracting user location data from the quality assessment database refers to analyzing the geographical distribution of customer sources. User location data can be obtained through various means, such as IP address resolution, mobile phone base station location, and membership registration information. Calculating the percentage of users visiting from provincial and municipal regions involves counting the number of customers from each province and city, dividing that number by the total number of customers, and obtaining the percentage value. This calculation considers both deep penetration within core business districts and coverage capabilities in broader markets. Geographical distribution data reflects the shopping mall's geographical reach and influence, which is crucial for assessing the accuracy and attractiveness of its positioning.

[0076] User consumption data is extracted from the quality assessment database, and consumption preferences are calculated by weighting consumption categories, spending levels, and purchase frequency. This process involves complex multi-dimensional data analysis, which can be represented by the following formula:

[0077]

[0078] Where CP represents consumption preference, T represents the total number of time periods, and ω t This represents the weighting coefficient for time period t. C represents the total number of consumer product categories, and ν... c CC represents the importance weight of category c. t,c TC represents the number of times product category c is consumed within a time period t. t This represents the total number of transactions within time period t. V represents the total number of spending levels, and ρ represents the total number of transactions. v CV represents the weighting coefficient of the monetary level v. t,v This represents the number of purchases at each monetary level (v) within a time period (t). F represents the number of purchase frequency groups, and σ represents the number of purchases at each monetary level (v). f CF represents the weighting coefficient of frequency group f. t,f This represents the number of consumers in frequency group f within time period t. This formula comprehensively considers three dimensions: product category, spending level, and purchase frequency, and calculates the user's consumption preference characteristics through weighted calculation. For example, for a shopping mall, analyzing the proportion of customer spending in different categories such as clothing, catering, and entertainment, as well as the distribution of high, medium, and low spending levels, and combining the characteristics of different frequency groups such as visiting once, 2-3 times, and more than 4 times per month, a consumer preference profile of the mall is formed through weighted calculation.

[0079] By combining device type distribution, visit time distribution, geographic distribution, and consumption preferences across multiple dimensions and establishing weighting coefficients, a site user profile is obtained. Multi-dimensional combination refers to integrating the feature data from these four aspects into a unified data structure to form a complete user behavior profile. The weighting coefficients are typically established based on the importance and discriminative power of the features and can be determined using methods such as expert experience, historical data analysis, or machine learning algorithms. The site user profile is a digital description of the characteristics of the shopping mall's customer group, including information from multiple dimensions such as device usage habits, visit time patterns, geographic origin distribution, and consumption behavior preferences.

[0080] By constructing target merchant customer profiles based on demographic characteristics, spending power levels, and product category preferences, we can obtain target merchant user profiles. Demographic characteristics include basic attributes such as age, gender, education level, and occupation; spending power levels reflect the economic strength and price sensitivity of the target customer group; and product category preferences describe the target customer group's interests in different categories of goods or services. Target merchant user profiles are digital descriptions of a merchant's ideal customer base, typically provided by the merchant based on its own business characteristics and market positioning, or generated by analyzing customer data from the merchant's other stores.

[0081] Feature vectors are calculated for corresponding dimensions of the site user profile and the target merchant user profile. The dimensional differences are then calculated using the Euclidean distance formula to obtain a matching score. Feature vector calculation involves converting the features of the two profiles into comparable numerical vectors. Euclidean distance is the straight-line distance between two points in n-dimensional space, calculated as the square root of the sum of the squares of the differences between the two points in each dimension. In user profile matching calculation, the smaller the Euclidean distance, the closer the two profiles are, and the higher the matching degree. Typically, the calculated distance values ​​are normalized or transformed into matching scores between 0 and 1; values ​​closer to 1 indicate a higher degree of matching.

[0082] For example, a well-known coffee brand plans to open a new store in a shopping mall and needs to assess whether the mall's customer base matches its target customer characteristics. First, analyzing the equipment type data in the mall's quality assessment database reveals that 88% of customers use mobile devices, 7% use wayfinding terminals, and 5% use membership cards, indicating a high level of mobile internet activity among the customer base. Next, analyzing the distribution of visit times identifies peak traffic periods as 12:00-14:00 and 17:30-20:00 on weekdays, and 14:00-18:00 on weekends, aligning with the work and rest patterns of office workers and families. Then, analyzing the geographical distribution determines that the main customer base comes from upscale residential areas and office building clusters within a 5-kilometer radius. Finally, analyzing consumption data and applying a weighted calculation formula, it is determined that customers in the mall spend a relatively high percentage on food and beverage (35%), with mid-to-high-end dining spending (average 100-300 RMB per person) dominating (60%), and purchases occurring 2-3 times per week (45%). These feature data were combined from multiple dimensions to form a user profile for the shopping mall. Meanwhile, the coffee brand provided its own target customer profile: primarily white-collar workers aged 25-45, with medium to high spending power, a preference for quality lifestyles, and main consumption times during weekday lunch breaks, after get off work, and weekend leisure periods. Feature vector calculations and Euclidean distance calculations were performed on the two profiles, yielding a matching score of 0.82 (out of 1), indicating a high degree of match between the shopping mall's customer characteristics and the coffee brand's target customers, making it suitable for opening a new store.

[0083] In one specific embodiment, the process of executing step S104 may specifically include the following steps:

[0084] (1) From the historical operating data of the shopping mall's traffic area, the number of customers is counted according to the daily time granularity to obtain the customer traffic data;

[0085] (2) Statistical analysis of user consumption behavior in the passenger flow data, and conversion rate data is obtained by calculating the ratio of the number of users to the passenger flow.

[0086] (3) Statistically analyze the user's spending amount in the conversion rate data, and calculate the average order value by the ratio of spending amount to the number of users.

[0087] (4) Group the customer flow, conversion rate and average transaction value at different times in the shopping mall to obtain business hours data;

[0088] (5) Extract promotional activity types, discount levels, and activity scope from the traffic flow areas of the shopping mall to obtain promotional activity format data;

[0089] (6) Combine the matching score with customer traffic data, conversion rate data, average order value data, business hours data and promotional activity data to obtain business characteristic data;

[0090] (7) Normalize the operational characteristic data and perform weighted calculations in conjunction with the site quality score to obtain the predicted score for site operation performance.

[0091] Specifically, in the mall traffic site selection method, operational performance prediction is a crucial step, directly impacting the accuracy of the final selection results. Statistical analysis of customer traffic from historical operational data of mall traffic sites, broken down by daily time granularity, involves extracting customer traffic records for different time periods (usually hourly) from raw data collected from channels such as customer count devices at entrances, camera recognition systems, or Wi-Fi probes, and then summarizing and statistically analyzing these records. Customer traffic refers to all visits to the mall within a specific time period, including repeat visits. This customer traffic data is organized by date and time period, forming a two-dimensional table. The horizontal axis represents the date, and the vertical axis represents the time period (e.g., 10:00-11:00, 11:00-12:00, etc.). The values ​​in the table represent the number of visits for the corresponding time unit. The consumption behavior of users within the customer traffic data is then statistically analyzed. The conversion rate is calculated by the ratio of the number of users making purchases to the total customer traffic. Consumption behavior statistics refer to extracting the number of users who actually made purchases from the mall's sales system, POS system, or membership management system. The number of paying customers refers to the number of customers who completed a purchase at any store within the mall during a specific time period. This number is obtained by integrating sales data from all stores within the mall. The conversion rate is calculated by dividing the number of paying customers by the customer traffic during the same period. It indicates the proportion of customers who entered the mall and actually made a purchase, and is an important indicator for measuring the effectiveness and attractiveness of the mall's marketing. Conversion rate data is also organized by date and time period, and has the same structure as customer traffic data.

[0092] The average transaction value (ACR) is calculated by statistically analyzing user spending amounts from conversion rate data and then dividing the ACR by the number of users. Spending amount statistics refer to the total amount of all sales transactions within a specific time period extracted from the mall's sales system. The ACR is calculated by dividing the total ACR by the number of users, representing the average spending per customer and reflecting the mall's customer spending power and product value positioning. ACR data is also organized by date and time period, forming a data table with the same structure as customer traffic and conversion rate. Customer traffic, conversion rate, and ACR are grouped according to different time periods within the mall's traffic flow area to obtain operating hour data. Data grouping refers to classifying and integrating time periods with similar characteristics based on the time distribution patterns of these three indicators. Common grouping methods include cluster analysis (such as the K-means algorithm) or experience-based divisions (such as morning market 9:00-12:00, lunch market 12:00-14:00, afternoon tea 14:00-17:00, evening market 17:00-21:00, etc.). The business hours data not only includes the time range of each group, but also the average customer traffic, average conversion rate, and average transaction value of each group, as well as statistical characteristics such as the coefficient of variation of these indicators, comprehensively describing the differences in the mall's operating performance at different times.

[0093] Promotional activity type data is obtained by extracting information on promotional activity types, discount levels, and activity scope from high-traffic areas within the shopping mall. Promotional activity types include different categories such as holiday promotions, member-exclusive activities, brand collaborations, and seasonal discounts. Discount level refers to the degree of price reduction during the promotion period, usually expressed as a discount percentage or a minimum spend amount. Activity scope refers to the coverage of participating stores and product categories. This promotional activity format data is obtained by collecting historical marketing plans from the mall, discount code usage records from the sales system, and merchant participation data. It is then structured and transformed into quantifiable feature vectors, such as one-hot encoding of the activity type, numerical representation of the discount level, and the coverage rate of the activity scope.

[0094] The matching score is combined with customer traffic data, conversion rate data, average transaction value data, business hours data, and promotional activity data to obtain operational characteristic data. The matching score refers to the degree of match between the mall's user profile calculated in the previous step and the target merchant's user profile. Feature combination refers to integrating these different dimensions of data into a unified data structure to form a complete description of operational characteristics. Combination methods include direct concatenation, weighted average, and principal component analysis, aiming to combine user matching with actual operational performance to form a more comprehensive evaluation basis. Operational characteristic data is a multi-dimensional vector or matrix, where each dimension represents an evaluation indicator, collectively describing the mall's operational potential and performance.

[0095] The operational characteristic data is normalized and then weighted according to the site quality score to obtain a predicted score for site operation performance. Normalization involves converting indicators with different dimensions and ranges to the same scale (usually a 0-1 range). Common methods include maximum / minimum standardization and Z-score standardization. The site quality score is a comprehensive score calculated in the previous steps based on the basic customer flow score, user stickiness coefficient, user participation coefficient, and site attribute coefficient. Weighted calculation involves assigning different weights to each indicator and then calculating a weighted average. The weighting is usually determined based on industry experience or historical data analysis. The predicted score for site operation performance is a value between 0 and 1; a higher value indicates greater potential for the shopping mall's operation.

[0096] For example, a shopping mall plans to introduce a children's education brand and needs to assess whether the mall's operating environment is suitable for the brand. First, daily foot traffic data is extracted from the mall's historical operating data. It is found that the average daily foot traffic is 15,000 from Monday to Friday and 25,000 from weekends, with peak traffic periods of 10:00-12:00 and 15:00-18:00. Next, consumption data is extracted from the sales system, calculating an average conversion rate of 35% from Monday to Friday and 45% on weekends, indicating that about one-third to one-half of visitors will make a purchase in the mall. Then, the average transaction value is calculated based on the amount spent and the number of customers. It is found that the average transaction value is 220 yuan from Monday to Friday and 320 yuan on weekends, reflecting an increased proportion of family spending on weekends. Grouping foot traffic, conversion rate, and average transaction value by time period reveals that 10:00-12:00 and 15:00-18:00 on weekends are the golden times for family spending, with sales of children's related products significantly higher during these times than at other times. Data extracted from promotional activity records revealed that promotions related to children, such as Children's Day and the back-to-school season, showed significantly higher conversion rates, and the mall's monthly parent-child activities generated a stable flow of target customers. The user profile matching score calculated in the previous steps (0.85, indicating a high degree of match between the mall's customer base and the educational brand's target users) was combined with the aforementioned operational data to form a complete operational characteristic description. Finally, these characteristic data were normalized and combined with the mall's site quality score (0.92, based on the calculations in the previous steps). A weighted calculation yielded a predicted site operation performance score of 0.88, indicating that the mall has high operational potential for children's educational brands and is suitable as a location for new stores.

[0097] In one specific embodiment, the process of executing step S105 may specifically include the following steps:

[0098] (1) Compare the predicted operating performance score with the industry average level, and conduct quality access assessment by setting a minimum access score to obtain a list of accessible sites;

[0099] (2) Based on the matching score of each shopping mall traffic venue in the list of approved venues, the user overlap is calculated to obtain the user matching evaluation result.

[0100] (3) By analyzing the historical operating performance of each shopping mall in the list of target venues, three levels of performance standards (high, medium and low) are set to obtain the expected performance classification data.

[0101] (4) Statistically analyze the average daily customer flow of shopping malls in the list of target sites, and set the upper and lower limits of site selection based on the merchant's investment budget to obtain the merchant site selection threshold.

[0102] (5) Cross-validate the expected effect classification data with the merchant site selection threshold, filter out the mall traffic site combinations that meet the conditions, and obtain the preliminary site list.

[0103] (6) The overlap of the population covered by each shopping mall in the initial selection list is analyzed, and the site selection combination results are obtained by deduplication.

[0104] Specifically, comparing the predicted operating performance score with the industry average is the first step in the screening process. The predicted operating performance score is a value between 0 and 1, obtained through multi-dimensional feature combination and weighted calculation in the previous steps, representing the mall's operating potential. The industry average is a benchmark value calculated using industry association data, market research reports, or historical data. The comparison calculation uses a simple difference method or ratio method to calculate the difference between the current mall and the industry benchmark. The minimum entry score is a set threshold value, usually 10%-20% higher than the industry average, to ensure that the selected malls have a competitive advantage. The quality entry assessment filters out malls with predicted operating performance scores higher than the minimum entry score, forming a list of eligible sites. This list includes basic information such as mall name, location, and area, as well as assessment data such as predicted operating performance scores and industry comparison results.

[0105] Based on the matching score of each shopping mall's traffic area in the list of approved venues, the user overlap is calculated to obtain the user matching evaluation result. The matching score is the similarity between the shopping mall user profile calculated in the previous steps and the target merchant's user profile. The calculation of user overlap involves more complex data analysis, which can be expressed by the following formula:

[0106]

[0107] Where UOI represents the User Overlap Index, κ represents the direct match weight coefficient (ranging from 0 to 1), MS represents the match score, S represents the total number of merchant categories in the mall, and η s TS represents the correlation coefficient of category s. s TS represents the trading share of category s. total This represents the total transaction share. The formula comprehensively considers both direct user profile matching and indirect related merchant transaction data, reflecting a more comprehensive user matching relationship. For example, when evaluating the location of a restaurant brand, it considers not only the direct matching between mall customers and the target customer group, but also the transaction situation of other related businesses within the mall, such as restaurants and entertainment venues, because these businesses often have a high degree of customer overlap with the target restaurant brand. The user matching evaluation result is a comprehensive indicator reflecting the overall fit between the mall's customer base and the merchant's target customers. By analyzing the historical operating performance of each mall's traffic locations in the target site list, a three-tiered performance standard (high, medium, and low) is set to obtain the expected performance grading data. Historical operating performance includes multiple indicators such as customer traffic growth rate, shop rent growth rate, sales growth rate, and brand turnover rate. This data is extracted from the mall management system, lease contract files, and sales reports, typically collecting data from the most recent 12-36 months. Analysis methods include trend analysis, seasonality analysis, and volatility analysis. The setting of the high, medium, and low performance standards usually adopts the quantile method or the natural breakpoint method, determining the dividing points based on the data distribution characteristics. The expected results grading data includes the scores of each shopping mall's various indicators and their comprehensive grading results, providing a stratification basis for subsequent screening.

[0108] The daily average foot traffic of shopping malls in the target site list is statistically analyzed. Based on the merchant's investment budget, upper and lower limits for site selection are set to obtain the merchant site selection threshold. Daily average foot traffic refers to the average number of visitors per day in a shopping mall within a complete operating cycle (e.g., one year), and is an important indicator of mall size. The calculation method is to divide the total number of visitors within a specific period by the number of operating days. The merchant's investment budget includes the sum of various costs such as rent, decoration, equipment, and initial operating capital. The determination of the upper and lower limits for site selection is based on an investment return model, considering factors such as industry average spending per customer, conversion rate, gross profit margin, cost structure, and target payback period. The merchant site selection threshold is a foot traffic range, ensuring that shopping malls within this range can meet the merchant's sales targets without negatively impacting the return on investment due to excessive costs.

[0109] Cross-validation was performed between the expected performance grading data and the merchant site selection threshold to filter out suitable combinations of shopping mall locations, resulting in a preliminary list of shortlisted locations. Cross-validation is a two-dimensional filtering method, with the horizontal axis representing the expected performance grading (high, medium, low) and the vertical axis representing whether the customer traffic falls within the threshold range (yes / no). Only shopping malls that simultaneously meet the criteria of "high or medium expected performance" and "customer traffic within the threshold range" are retained. This cross-validation ensures that the selection results have both good operational potential and are suitable for the merchant's investment scale. The preliminary list of shortlisted locations contains information on shopping malls that meet both criteria, and its number is significantly less than the list of eligible locations, making it closer to the final recommended results.

[0110] The overlap analysis of the target customer groups of each shopping mall in the initial shortlist is performed, and deduplication is used to obtain the final site selection combination. The target customer group refers to the shopping mall's customer base, typically determined by its geographical location, accessibility, and catchment area. The overlap analysis uses Geographic Information System (GIS) technology to calculate the overlapping area of ​​different shopping mall catchment areas and the population within those areas. Deduplication employs a set covering problem approach, aiming to select the minimum number of shopping malls to cover the maximum target customer area. Common algorithms include greedy algorithms, genetic algorithms, or integer programming. The final site selection combination result is the recommended site selection plan for merchants, including all relevant information about the recommended shopping malls and an analysis of the combination's advantages.

[0111] For example, a chain clothing brand plans to open a new store in a region and first collects data from 30 candidate shopping malls. Analysis of their predicted operating performance scores reveals that these malls' scores range from 0.65 to 0.92, while the industry average is 0.75. A minimum admission score of 0.80 is set, and after screening, 18 malls are selected for the shortlist. Next, the user overlap index for these 18 malls is calculated, with the direct matching coefficient κ set at 0.6, and the correlation coefficient η for each merchant category set as follows: similar clothing brands 0.8, accessory brands 0.6, catering brands 0.4, and other brands 0.2. Analysis of the transaction share of different categories of merchants within each mall shows that the user overlap index for the 18 malls ranges from 0.55 to 0.88. Based on merchant requirements, a user overlap index greater than 0.70 is set, and after further screening, 12 malls are selected. Then, the historical operating data of these 12 shopping malls over the past 24 months were analyzed, including indicators such as customer traffic growth rate (mean 17.5%, standard deviation 6.8%), rent growth rate (mean 8.2%, standard deviation 2.5%), and sales growth rate (mean 15.3%, standard deviation 5.4%). A comprehensive scoring method was used to calculate the overall performance, and the malls were divided into three categories based on the natural breakpoint method: high-performance (4 malls, score > 85), medium-performance (6 malls, score 70-85), and low-performance (2 malls, score < 70). Simultaneously, based on the apparel brand's single-store investment budget of 4.5 million yuan and target payback period of 18 months, combined with industry average transaction value per customer of 350 yuan, conversion rate of 15%, and gross profit margin of 60%, the appropriate daily customer traffic for a shopping mall was calculated to be between 6,000 and 12,000 visits. The expected performance classification was cross-validated with the customer traffic threshold, ultimately retaining 3 malls with high performance plus the threshold, and 4 malls with medium performance plus the threshold, for a total of 7 malls, into the initial site selection list. Finally, an analysis of the geographical locations and catchment areas of these seven shopping malls revealed significant overlap in their customer bases. An overlap matrix was calculated using a GIS system, and a greedy algorithm was employed for optimization. Ultimately, the three malls with the widest coverage and minimal overlap were selected as the final recommendations, forming the site selection combination.

[0112] In one specific embodiment, the process of executing step S106 may specifically include the following steps:

[0113] (1) Real-time monitoring of the quality assessment indicators of each shopping mall traffic site in the site selection combination results, and quality fluctuation data are obtained by calculating the deviation with historical benchmark values.

[0114] (2) Based on the matching degree of user profiles of shopping mall traffic venues in the venue screening and combination results, the user group characteristics are regularly updated and compared to obtain user change data.

[0115] (3) Collect the operational data of the shopping mall traffic venues in the venue selection combination results, and obtain the actual operational performance data by calculating the difference between the actual conversion rate and the expected conversion rate.

[0116] (4) By comprehensively analyzing the quality fluctuation data, user change data and actual operating effect data, and making early warning judgments according to the set thresholds, site early warning information is obtained.

[0117] (5) Based on the site warning information, the traffic flow of shopping mall sites with different warning levels is classified and processed, and the effect comparison results are obtained by comparing and analyzing with historical performance.

[0118] (6) The results of the effect comparison are dynamically scored according to the weight indicators, and the traffic flow of the shopping mall is reordered and replaced to obtain the shopping mall site optimization plan.

[0119] Specifically, after the initial screening, the mall traffic site selection method needs to establish a dynamic optimization mechanism to ensure the continued effectiveness of the recommendation results. Real-time monitoring of the quality assessment indicators for each mall traffic site in the site selection combination results refers to continuously collecting key indicator data from the selected malls and performing dynamic analysis. Quality assessment indicators include core indicators calculated in previous steps, such as basic customer flow score, user stickiness coefficient, user participation coefficient, and site attribute coefficient. Real-time monitoring is typically conducted daily or weekly to ensure timely capture of indicator changes. Deviation calculation by comparing with historical benchmark values ​​involves comparing the current monitored value with the mall's average performance over a past period (usually 3-6 months) or its performance in a specific period (such as the same period last year). Deviation calculation methods include the simple difference method (current value minus benchmark value), the percentage difference method (difference divided by benchmark value), or the standard deviation multiple method (difference divided by historical standard deviation). Quality fluctuation data is a structured dataset containing information such as the current value, benchmark value, deviation value, and trend of each indicator, reflecting the dynamic changes in the mall's quality status. Based on the user profile matching degree of the shopping mall traffic venues in the venue selection combination results, user group characteristics are regularly updated and compared to obtain user change data. User profile matching degree refers to the similarity between the shopping mall user profile and the target merchant's user profile. User group characteristics include multiple dimensions such as device type distribution, access time distribution, geographical distribution, and consumption preferences. Regular update and comparison refers to re-collecting user data monthly or quarterly, updating user profiles, and comparing and analyzing them with previous profiles. Comparison methods include feature vector difference calculation, distribution similarity measurement (such as KL divergence, JS divergence), or time-series comparison of specific indicators. User change data records the changes in each feature dimension, including information such as the magnitude, direction, and rate of change, reflecting the evolution trend of the shopping mall's customer structure.

[0120] The operational data of shopping mall locations selected from the site screening results are collected. Actual operational performance data is obtained by calculating the difference between the actual conversion rate and the expected conversion rate. Operational data includes commercial indicators such as customer traffic, transaction volume, and sales revenue, collected through the mall's sales system, POS system, or third-party monitoring tools. The conversion rate, the proportion of actual customers making a purchase to total customer traffic, is a core indicator for measuring mall operational performance. The expected conversion rate is a target value predicted based on historical data and market conditions, typically determined during the screening phase. The difference is calculated using simple subtraction or the percentage deviation method to quantify the gap between actual performance and expectations. The actual operational performance data includes the actual values, expected values, differences, and achievement rates of various operational indicators, comprehensively reflecting the mall's operational performance.

[0121] By comprehensively analyzing quality fluctuation data, user change data, and actual operational performance data, and making early warning judgments based on set thresholds, site early warning information is obtained. Comprehensive analysis refers to integrating the three types of data to find interrelationships and influencing factors. Analytical methods include correlation analysis, factor analysis, or regression analysis. Setting thresholds refers to setting early warning trigger points for various indicators, usually based on industry experience or historical data statistics. Common threshold setting methods include fixed value methods (e.g., triggering an early warning when the conversion rate drops by more than 10%), percentile methods (e.g., triggering an early warning when the indicator falls below the historical 25th percentile), or multiple condition methods (e.g., triggering an early warning only when multiple conditions are met simultaneously). Early warning judgment determines the early warning status and level based on whether the indicator exceeds the threshold. Site early warning information includes the warning site, warning indicator, warning level, trigger time, and specific manifestations, providing a basis for subsequent adjustments.

[0122] Based on site alert information, shopping mall locations with different alert levels are categorized and processed. The results are then compared with historical performance to obtain effectiveness comparison results. Categorization involves employing different analysis strategies based on the alert level and type. Common categorizations include red alert (significantly deviating from expectations, requiring immediate adjustment), yellow alert (significantly deviating from expectations, requiring close monitoring), and blue alert (slightly deviating from expectations, optimization recommended). Historical performance comparison analysis involves comparing the current performance of alerted locations with their past performance over multiple time periods, and also combining this with industry averages and leading site performance for horizontal comparison. The effectiveness comparison results are a comprehensive analysis report, including detailed data comparisons, analysis of the causes of problems, and assessment of improvement potential.

[0123] The results of the performance comparison are dynamically scored according to weighted indicators. This process is used to re-rank and replace shopping mall locations, resulting in a mall location optimization plan. The weighted indicators are determined based on the merchants' strategic priorities and the market environment, typically including multiple dimensions such as quality stability, customer base matching, operational performance, and development potential. Dynamic scoring involves assigning scores to each shopping mall on each weighted indicator based on the latest data, and then calculating a weighted total score. Re-ranking involves adjusting the priority order of existing shopping malls based on the dynamic scoring results. Replacement involves considering replacing warning locations with alternative locations when the score of a warning location is significantly lower than that of a backup location. The mall location optimization plan is the final adjustment recommendation, including complete content such as retaining locations, replacing locations, adjustment strategies, and expected results.

[0124] For example, a mobile phone brand selected three shopping malls in a region as locations for new stores. After six months of operation, dynamic monitoring and optimization were implemented. First, the quality assessment indicators of these three malls were monitored in real time. Raw data such as customer flow, dwell time, and bounce rate were collected weekly to calculate basic customer flow scores, user stickiness coefficients, and other quality indicators. These indicators were compared with baseline values ​​from the three months prior to opening. It was found that the user stickiness coefficient of mall A dropped from 1.15 to 0.92, a decrease of 20%, exceeding the 15% warning threshold; while the changes in indicators for malls B and C were within reasonable ranges. Next, the user profiles of each mall were updated monthly, and the changes between the current profile and the initial profile were compared using the feature vector difference calculation method. The user profile of mall A changed significantly, especially the distribution of visiting times shifted from mainly afternoon to mainly evening, and the consumption preference shifted from mainly mid-to-high-end electronic products to mainly fast-moving consumer goods; the customer structure of mall B remained basically stable; the geographic distribution of users in mall C expanded, with the coverage area increasing by approximately 20%. Simultaneously, sales data from the three malls were collected to calculate the difference between the actual conversion rate and the expected conversion rate. Mall A's actual conversion rate was 8%, far below the expected 15%; Mall B's was 14%, close to the expected 15%; and Mall C's was 18%, exceeding the expected 15%. Analyzing these three types of data and comparing them to the preset warning thresholds, Mall A triggered a red warning (multiple core indicators significantly deviated from expectations), Mall B was in normal condition, and Mall C performed better than expected. Based on the warning information, a deeper analysis of Mall A revealed that the problem began four months ago with a business format adjustment, where a large number of catering brands changed the customer flow structure. Simultaneously, an investigation of Mall D, a candidate mall, revealed its recent excellent performance, with its user profile highly matching the target customer group of the mobile phone brand. Finally, a dynamic scoring was conducted using weighted indicators of "customer group matching 40%, operating performance 30%, development potential 20%, and location complementarity 10%." Mall A scored 62, significantly lower than Mall D's 86. Therefore, the optimization plan recommends replacing Mall A with Mall D, while retaining the well-performing Malls B and C, forming a new venue combination.

[0125] In one specific embodiment, the process of performing a comprehensive analysis of quality fluctuation data, user change data, and actual operating performance data may specifically include the following steps:

[0126] (1) The quality fluctuation data are grouped and statistically analyzed according to the daily fluctuation amplitude. By calculating the number of consecutive fluctuation days and the direction of fluctuation, the quality fluctuation trend data is obtained.

[0127] (2) Decompose the user change data according to the change dimensions, and obtain the user change trend data by calculating the change magnitude and change rate of each dimension;

[0128] (3) Perform time series analysis on the actual operating performance data, and obtain the performance trend data by calculating the cumulative value and slope of the performance deviation;

[0129] (4) Compare the quality fluctuation trend data, user change trend data and effect change trend data in the same dimension, and obtain trend correlation data through correlation analysis;

[0130] (5) Quantitatively calculate the degree of abnormality of each indicator in the trend correlation data, and obtain the warning level data by comparing it with the preset warning threshold;

[0131] (6) Associate and mark the warning level data with abnormal indicators, and obtain site warning information through summary analysis.

[0132] Specifically, the early warning mechanism in the mall traffic screening method requires a refined data processing workflow to ensure the accuracy and timeliness of monitoring. Grouping and statistically analyzing quality fluctuation data according to daily fluctuation amplitude is a fundamental step in early warning judgment. Daily fluctuation amplitude refers to the difference between each quality indicator and the benchmark value on that day, usually expressed as a percentage. Grouping statistics use a numerical range division method, classifying fluctuation amplitude into multiple levels, such as slight fluctuation (0-5%), moderate fluctuation (5-10%), significant fluctuation (10-20%), and severe fluctuation (>20%). Calculating the number of consecutive fluctuation days refers to the number of days a specific indicator continuously changes in the same direction (upward or downward). The fluctuation direction records the trend of indicator change. For example, if customer traffic decreases for 7 consecutive days, and the fluctuation amplitude exceeds 10% each time, it is recorded as a fluctuation record of "customer traffic - significant - 7 days". Quality fluctuation trend data is a structured information set containing the distribution of fluctuation levels, duration, and direction information of each indicator, which can intuitively reflect the dynamic changing trend of the mall's quality status.

[0133] Decomposing user change data according to change dimensions is a key step in gaining a deeper understanding of changes in customer structure. Change dimensions include multiple aspects such as device type distribution, access time distribution, geographic distribution, and consumption preferences. The decomposition process breaks down changes in the overall user profile into specific changes in each dimension. The magnitude of change in each dimension is calculated—the difference between the current value and the baseline value—usually quantified using mathematical tools such as Euclidean distance or Mahalanobis distance. The rate of change refers to the magnitude of change divided by the time interval, reflecting the speed of change. For example, if a shopping mall's consumption preference shifts from primarily high-end electronics to primarily mid-range clothing within three months, a change of 40%, the rate of change is 13.3% / month. This rate is significantly higher than normal market seasonal changes (usually around 5% / month). User change trend data records the specific details of changes in each dimension, including the main direction of change, key change indicators, and change rate rankings, providing a panoramic view of customer group changes for subsequent analysis.

[0134] Time series analysis of actual operational performance data is a scientific method for assessing the performance trend of a shopping mall. Time series analysis uses statistical tools such as moving averages, exponential smoothing, or ARIMA models to identify trend, seasonal, and random components in the operational data. By calculating the cumulative value of performance deviation—the sum of deviations at each time point—long-term performance gaps can be assessed. For example, if a shopping mall's conversion rate is lower than expected for three consecutive months, with monthly deviations of -2%, -3%, and -4% respectively, and a cumulative deviation of -9%, this continuous decline and accelerating deterioration trend is more concerning than a single large fluctuation. The slope of change refers to the slope of the trend line after linear regression fitting, visually showing the speed and direction of indicator changes over time. Performance trend data includes trend parameters, cyclical characteristics, and outlier information for each operational indicator, providing a complete time-dimensional insight for shopping mall performance evaluation.

[0135] Comparing quality fluctuation trends, user change trends, and performance change trends within the same timeframe is a crucial method for identifying the root causes of problems. Same-dimensional comparison refers to comparing these three types of trend data side-by-side on the same time scale to identify the sequence of changes and their mutual influence. Correlation analysis methods, such as Pearson correlation coefficient, Spearman rank correlation, or cross-lag correlation, are used to calculate the strength of the association between different indicators. For example, analysis might reveal that changes in a shopping mall's user profile (decreased spending power) lead changes in quality indicators (decreased user stickiness) by 1-2 weeks, while changes in quality indicators lead changes in performance (decreased conversion rate) by 1-2 weeks. This temporal sequence reveals a potential causal chain. Trend correlation data records the correlation coefficients, lead-lag relationships, and possible causal paths between each indicator, providing a systematic analytical foundation for early warning judgments.

[0136] Quantifying the degree of anomaly in each indicator within trend-related data is the core step in early warning judgment. Anomaly quantification employs various statistical methods, such as the Z-score method (the difference between the current value and the mean divided by the standard deviation), anomaly detection algorithms (such as IQR and DBSCAN clustering), or rule-based scoring systems. By comparing the result with preset early warning thresholds, it is determined whether an indicator triggers an early warning. Preset early warning thresholds are determined based on industry experience, historical data analysis, or statistical significance principles. Common settings include a Z-score exceeding 2 (approximately 95% confidence level), a deviation exceeding 30%, or a continuous decline for three weeks. Early warning levels are typically divided into multiple levels, such as Level 1 (red, requiring immediate action), Level 2 (yellow, requiring close monitoring), and Level 3 (blue, optimization recommended), each level corresponding to different degrees of anomaly and urgency.

[0137] Associating and tagging warning level data with abnormal indicators is the final step in forming complete warning information. Association tagging binds the warning level with specific abnormal indicators, causes of the anomaly, and scope of impact, creating a structured warning record. Through summary analysis, all warning information is integrated and prioritized to obtain the final site warning information. Summary analysis methods include decision tree analysis, expert system evaluation, or comprehensive scoring mechanisms. For example, when a shopping mall simultaneously triggers both a user profile change warning and a business performance warning, and the two are highly correlated, the warning level will be raised and tagged as "business decline due to changes in customer structure." This correlation analysis provides more valuable decision-making information than individual indicator warnings. Site warning information is a comprehensive warning report, including the warning site, warning level, key abnormal indicators, possible causes, scope of impact, recommended measures, and follow-up points, providing specific operational guidance for the mall's dynamic optimization and adjustment.

[0138] For example, after a shopping mall was selected as the location for a high-end cosmetics brand's store, it entered a dynamic monitoring phase. First, daily data on quality indicators such as customer traffic and dwell time were collected, and the deviations from baseline values ​​were calculated and grouped according to the magnitude of the deviation. After three months of continuous monitoring, it was found that the dwell time indicator began to decline continuously from the eighth week, dropping from the baseline of 125 minutes to 110 minutes, remaining within a deviation range of approximately -10% for 12 consecutive days, forming a quality fluctuation trend record of "dwell time - moderate decline - 12 days". Simultaneously, user profile data was updated monthly, and the changes in each dimension were broken down. The most significant change was found in the age distribution dimension, with the proportion of people under 35 years old decreasing from 60% to 45%, a 25% change over three months, with an average monthly change rate of approximately 8.3%, far exceeding the normal fluctuation level of 3%, recorded as a rapid change; consumption preferences also shifted from primarily cosmetics and clothing to primarily dining and family-oriented items, a change of approximately 30%. Operational data analysis shows that the conversion rate of high-end goods in the mall has been declining continuously from a benchmark of 8% to 5.5%, with a cumulative deviation of -8.5%. The slope of the change after linear regression is -0.83% / month, indicating a clear downward trend that has not yet bottomed out. Comparison and correlation analysis of the three types of trend data within the same dimension revealed a high correlation between the decrease in the young customer base and the decrease in dwell time (correlation coefficient 0.86). Furthermore, the decrease in the young customer base leads the decrease in dwell time by approximately two weeks, while the decrease in dwell time leads the decline in the high-end conversion rate by approximately one week, forming a clear chain of correlation. Quantifying the degree of anomaly in the relevant indicators, the Z-score for changes in the proportion of young customers is -2.4, the Z-score for changes in dwell time is -1.8, and the Z-score for changes in the high-end conversion rate is -2.2, all exceeding the preset warning threshold of ±1.5. Based on the degree of anomaly, a Level 1 warning (red) is triggered for the proportion of young customers, a Level 2 warning (yellow) for dwell time, and a Level 1 warning (red) for the high-end conversion rate. By linking the warning level with specific abnormal indicators and conducting comprehensive analysis, the core problem was identified as "the loss of young, high-spending customers leading to a deterioration in core operating indicators." Further analysis revealed that newly opened integrated entertainment centers in the vicinity were significantly diverting young customers. This resulted in a complete site warning system, including detailed data analysis, problem diagnosis, and adjustment suggestions, providing data support for subsequent site optimization decisions.

[0139] The above describes the traffic venue filtering method for marketing in the embodiments of this application. The following describes the traffic venue filtering system for marketing in the embodiments of this application. Please refer to [link / reference]. Figure 2 One embodiment of the traffic venue screening system for marketing in this application includes:

[0140] The data collection module is used to collect multi-dimensional data on the total number of visitors, number of unique visitors, dwell time and bounce rate of shopping mall traffic areas by deploying a site monitoring system. The data is tagged and stored according to the source channel, device and region to obtain an initial evaluation matrix.

[0141] The calculation module is used to calculate the basic passenger flow score, user stickiness coefficient, user participation coefficient and site attribute coefficient based on the initial evaluation matrix, and obtain the site quality evaluation database through weighted calculation.

[0142] The extraction module is used to extract user device type distribution, access time distribution, geographical distribution and consumption preference characteristics based on the quality assessment database, establish site user profiles, and calculate the matching degree score of target merchant user profiles;

[0143] The prediction module is used to obtain a predicted score for the venue's operating performance based on the matching score, combined with features such as customer traffic, conversion rate, and consumption amount in historical operating data, as well as venue quality score, business hours, and promotional activities.

[0144] The evaluation module is used to conduct quality access assessment and user matching assessment of the venue based on the predicted operating performance score, set grading standards according to the expected results, and filter the venues by combining the merchant site selection threshold to obtain the venue selection combination results.

[0145] The monitoring module is used to monitor and analyze the quality assessment indicators, user profile matching degree and actual operating effect of the site based on the site selection and combination results, and make dynamic adjustments through early warning triggering and effect comparison to obtain the shopping mall site optimization plan.

[0146] Through the collaborative efforts of the aforementioned components and the deployment of a site monitoring system for multi-dimensional data collection, comprehensive monitoring of key indicators such as total mall visitors, unique visitors, dwell time, and bounce rate was achieved. This approach overcomes the limitations of traditional methods that focus solely on single-dimensional customer flow. Furthermore, by employing tagging and storage technology, attributes such as source channels, devices, and regions were structurally organized to form an information-rich initial evaluation matrix, laying a solid data foundation for subsequent analysis. Based on this evaluation matrix, the method introduced multi-dimensional evaluation indicators such as basic customer flow scores, user stickiness coefficients, user engagement coefficients, and site attribute coefficients. Weighted calculations were used to generate a site quality evaluation database, enabling the scientific quantification of mall quality. Further, user device type distribution, visit time distribution, geographical distribution, and consumption preference characteristics were extracted from the quality evaluation database to establish a three-dimensional site user profile. This profile was then used to calculate and match user data using intelligent matching algorithms. The matching score of target merchant user profiles effectively solves the problem of traditional methods' difficulty in accurately assessing customer group matching. Based on this, and combining characteristics such as customer traffic, conversion rate, and spending in historical operating data, as well as factors such as site quality rating, operating hours, and promotional activities, a predictive model is applied to generate a predicted score for site operating performance, significantly improving the scientific rigor and accuracy of site selection decisions. Based on the predicted scores, the method designs a dual screening mechanism of quality access assessment and user matching assessment, and introduces expected effect grading standards and merchant site selection magnitude thresholds as constraints to obtain the optimal site selection combination. Finally, by establishing a monitoring and analysis closed loop, the quality assessment indicators, user profile matching degree, and actual operating performance of the selected sites are continuously monitored. Dynamic adjustments are made based on early warning triggering mechanisms and effect comparison analysis to generate mall site optimization plans, effectively addressing the challenges brought about by changes in the market environment and customer flow characteristics. The entire solution fully utilizes artificial intelligence algorithms for in-depth data mining and pattern recognition. In particular, advanced algorithms such as cluster analysis, similarity calculation, and regression prediction are introduced in the user profile construction, matching degree calculation, and business performance prediction stages. These algorithmic features play a key role in improving data processing accuracy, discovering potential correlation patterns, and achieving accurate prediction, significantly enhancing the efficiency and accuracy of shopping mall traffic site selection.

[0147] An embodiment of the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method. It is understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0148] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the present invention and embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual-rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM, etc.

[0149] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0150] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0151] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for filtering traffic venues for marketing, characterized in that, The method for filtering traffic venues used for marketing includes: By deploying a site monitoring system, multi-dimensional data collection is carried out on the total number of visitors, number of unique visitors, dwell time and bounce rate of shopping mall traffic areas. The data is tagged and stored according to the source channel, device and region to obtain an initial evaluation matrix. Based on the initial evaluation matrix, the basic visitor flow score, user stickiness coefficient, user engagement coefficient, and venue attribute coefficient are calculated. A venue quality evaluation database is obtained through weighted calculation, including: normalizing the total visitor data and unique visitor data in the initial evaluation matrix and linearly combining them in a 6:4 ratio to obtain the basic visitor flow score; converting the dwell time data in the initial evaluation matrix to minutes and calculating the user stickiness coefficient based on its ratio to the industry average dwell time; and performing a reverse conversion on the bounce rate data in the initial evaluation matrix and calculating the deviation from the industry benchmark bounce rate to obtain the user engagement coefficient. The system performs the following steps: First, based on the source channel data in the initial evaluation matrix, a weighted credibility score is assigned to each channel to obtain the source channel credibility. Second, based on the access device data in the initial evaluation matrix, the number and distribution ratio of covered device types are statistically analyzed to obtain the device coverage area. Third, based on the user geographic data in the initial evaluation matrix, the concentration and coverage breadth of geographic distribution are calculated to obtain the geographic distribution. Fourth, the source channel credibility, device coverage area, and geographic distribution are linearly combined to obtain the site attribute coefficients. Finally, the basic passenger flow score, user stickiness coefficient, user participation coefficient, and site attribute coefficients are weighted and summed to obtain the site quality assessment database. Based on the quality assessment database, user device type distribution, access time distribution, geographical distribution and consumption preference characteristics are extracted to establish site user profiles and calculate the matching score of target merchant user profiles. Based on the matching score, combined with historical operating data such as customer traffic, conversion rate, and spending amount, as well as venue quality score, operating hours, and promotional activities, a predicted score for venue operating performance is obtained. Based on the predicted operating performance score, the site is assessed for quality access and user matching degree. A grading standard is set according to the expected effect, and the site selection is screened in combination with the merchant site selection threshold to obtain the site selection combination result. Based on the site selection and combination results, the quality assessment indicators of the sites, the matching degree of user profiles, and the actual operating results are monitored and analyzed. Dynamic adjustments are made through early warning triggering and effect comparison to obtain the shopping mall site optimization plan.

2. The method for screening traffic venues for marketing according to claim 1, characterized in that, The system collects multi-dimensional data on total visitor volume, unique visitors, dwell time, and bounce rate in the shopping mall by deploying a site monitoring system. This data is then tagged and stored according to source channel, device, and region to obtain an initial evaluation matrix, including: The total number of visitors to each shopping mall location is obtained by collecting customer flow records for each location within a specified time window through a monitoring system. The user identification information in the total number of visitors to the shopping mall's traffic area is deduplicated to obtain the number of unique visitors to the shopping mall's traffic area. By recording the entry and exit times of users in shopping mall traffic areas, the time difference of each visit is calculated to obtain the dwell time in shopping mall traffic areas; The bounce rate of the shopping mall traffic area is calculated by comparing the number of users who only accessed a single area with the total number of users who accessed the area. Extract source channel identifiers based on user access information, and categorize them into online traffic, surrounding communities, business cooperation, and brand activities to obtain source channel data for shopping mall traffic venues; By parsing the user's device identification information, and classifying and statistically analyzing them according to mobile terminals, mall wayfinding devices, and membership cards, we can obtain access device data for mall traffic areas. Geographical location data is analyzed based on user location information, and distribution statistics are performed according to province, city, and region to obtain user geographic data for shopping mall traffic areas. By linking the source channel data, access device data, and user geographic data of the shopping mall traffic venue in multiple dimensions, a tag index system is established to obtain an initial evaluation matrix.

3. The method for screening traffic venues for marketing according to claim 1, characterized in that, The process involves extracting user device type distribution, access time distribution, geographic distribution, and consumption preference characteristics from the quality assessment database to establish a site user profile and calculating the matching score of the target merchant user profile, including: The device type data of users in the shopping mall traffic area is extracted from the quality assessment database, and the distribution ratio of mobile devices, wayfinding terminals and membership cards is statistically calculated to obtain the device type distribution. User access time data is extracted from the quality assessment database, and the time intervals are divided into 24-hour intervals. The access frequency of each time interval is statistically analyzed to obtain the access time distribution. Data on the user's location is extracted from the quality assessment database, and the proportion of users accessing the platform in provincial and municipal regions is calculated to obtain the geographical distribution. User consumption data is extracted from the quality assessment database, and consumption preferences are obtained by weighting the consumption categories, spending levels, and purchase frequency. The distribution of device type, access time, geographic distribution, and consumption preferences are combined in multiple dimensions, and weighting coefficients are established to obtain a site user profile. The target merchant's customer profile is constructed by using demographic features, consumption capacity levels, and category preferences. Feature vectors are calculated for the corresponding dimensions of the venue user profile and the target merchant user profile. The dimensional difference value is calculated using the Euclidean distance formula to obtain the matching score.

4. The method for screening traffic venues for marketing according to claim 1, characterized in that, The method of obtaining a predicted venue operating performance score based on the matching score, combined with historical operating data such as customer traffic, conversion rate, and spending amount, as well as venue quality rating, operating hours, and promotional activity formats, includes: Customer traffic data is obtained by statistically analyzing the number of customers in the shopping mall's historical traffic data according to the daily time granularity. The consumption behavior of users in the aforementioned customer flow data is statistically analyzed, and the conversion rate data is obtained by calculating the ratio of the number of consuming users to the customer flow. The average order value is calculated by statistically analyzing the user spending amount in the conversion rate data and then calculating the ratio of spending amount to the number of users. The customer traffic, conversion rate, and average transaction value at different times in the shopping mall are grouped to obtain business hour data; From the traffic flow areas of the shopping mall, the types of promotional activities, discount levels, and activity scope are extracted to obtain promotional activity format data; The matching score is combined with customer traffic data, conversion rate data, average order value data, business hours data, and promotional activity data to obtain operational characteristic data; The operational characteristic data is normalized and weighted by combining it with the site quality score to obtain the predicted score for site operation performance.

5. The method for screening traffic venues for marketing according to claim 1, characterized in that, Based on the predicted operating performance score, the site undergoes quality access assessment and user matching assessment. A grading standard is set according to the expected results, and a selection process is conducted using a merchant site selection threshold to obtain the site selection combination results, including: The predicted operating performance score is compared with the industry average. A quality access assessment is conducted by setting a minimum access score to obtain a list of accessible sites. Based on the matching score of each shopping mall traffic venue in the list of approved venues, the user overlap is calculated to obtain the user matching evaluation result. By analyzing the historical operating performance of each shopping mall in the list of approved venues, three levels of performance standards (high, medium, and low) were set to obtain expected performance classification data. The average daily customer flow of each shopping mall in the list of permitted sites is statistically analyzed, and the upper and lower limits of the site selection level are set according to the merchant's investment budget to obtain the merchant site selection level threshold. The expected effect classification data is cross-validated with the merchant site selection volume threshold to filter out the mall traffic site combinations that meet the conditions and obtain the preliminary site list. The overlap of the population covered by each shopping mall in the initial list of selected sites is analyzed, and the site selection combination results are obtained by deduplication.

6. The method for screening traffic venues for marketing according to claim 1, characterized in that, Based on the site selection and combination results, the quality assessment indicators, user profile matching degree, and actual operating effect of the sites are monitored and analyzed. Dynamic adjustments are made through early warning triggering and effect comparison to obtain a shopping mall site optimization plan, including: The quality assessment indicators of each shopping mall traffic venue in the venue selection combination results are monitored in real time, and the quality fluctuation data are obtained by calculating the deviation with the historical benchmark value. Based on the matching degree of user profiles of shopping mall traffic venues in the venue screening and combination results, the user group characteristics are regularly updated and compared to obtain user change data. The operational data of shopping mall traffic venues in the venue selection combination results are collected, and the actual operating effect data is obtained by calculating the difference between the actual conversion rate and the expected conversion rate. By comprehensively analyzing the quality fluctuation data, user change data, and actual operating performance data, and making early warning judgments according to the set thresholds, site early warning information is obtained. Based on the site warning information, shopping mall traffic sites with different warning levels are classified and processed, and the effect comparison results are obtained by comparing and analyzing with historical performance. The results of the effect comparison are dynamically scored according to weighted indicators, and the traffic flow areas of the shopping mall are reordered and replaced to obtain the shopping mall site optimization plan.

7. The method for screening traffic venues for marketing according to claim 6, characterized in that, The process involves comprehensively analyzing the quality fluctuation data, user change data, and actual operating performance data, and making early warning judgments based on set thresholds to obtain site early warning information, including: The quality fluctuation data is grouped and statistically analyzed according to the daily fluctuation amplitude. By calculating the number of consecutive fluctuation days and the fluctuation direction, the quality fluctuation trend data is obtained. The user change data is decomposed according to the change dimensions, and the user change trend data is obtained by calculating the change magnitude and change rate of each dimension. Time series analysis was performed on the actual operating performance data, and the cumulative value and slope of performance deviation were calculated to obtain performance trend data. The quality fluctuation trend data, user change trend data, and effect change trend data are compared in the same dimension, and trend correlation data are obtained through correlation analysis. The degree of abnormality of each indicator in the trend correlation data is quantitatively calculated, and the warning level data is obtained by comparing it with the preset warning threshold. The warning level data is associated with and marked with abnormal indicators, and the site warning information is obtained through summary analysis.

8. A traffic venue screening system for marketing, used to implement the traffic venue screening method for marketing as described in any one of claims 1-7, characterized in that, The traffic venue screening system used for marketing includes: The data collection module is used to collect multi-dimensional data on the total number of visitors, number of unique visitors, dwell time and bounce rate of shopping mall traffic areas by deploying a site monitoring system. The data is tagged and stored according to the source channel, device and region to obtain an initial evaluation matrix. The calculation module is used to calculate the basic visitor flow score, user stickiness coefficient, user engagement coefficient, and site attribute coefficient based on the initial evaluation matrix, and to obtain a site quality evaluation database through weighted calculation. This includes: normalizing the total visitor data and unique visitor data in the initial evaluation matrix and linearly combining them in a 6:4 ratio to obtain the basic visitor flow score; converting the dwell time data in the initial evaluation matrix to minutes and calculating the user stickiness coefficient based on the ratio to the industry average dwell time; and performing a reverse conversion on the bounce rate data in the initial evaluation matrix and calculating the deviation from the industry benchmark bounce rate to obtain the user stickiness coefficient. The evaluation process includes: 1) an engagement coefficient; 2) weighted credibility scores for each channel based on the source channel data in the initial evaluation matrix to obtain the source channel credibility; 3) statistical analysis of the number and distribution ratio of covered device types based on the access device data in the initial evaluation matrix to obtain the device coverage range; 4) calculation of the concentration and coverage breadth of the geographical distribution based on the user geographic data in the initial evaluation matrix to obtain the geographic distribution; 5) linear combination of the source channel credibility, device coverage range, and geographic distribution to obtain the site attribute coefficient; and 6) weighted summation of the basic passenger flow score, user stickiness coefficient, user engagement coefficient, and site attribute coefficient to obtain the site quality evaluation database. The extraction module is used to extract user device type distribution, access time distribution, geographical distribution and consumption preference characteristics based on the quality assessment database, establish site user profiles, and calculate the matching degree score of target merchant user profiles; The prediction module is used to obtain a predicted score for the venue's operating performance based on the matching score, combined with historical operating data such as customer traffic, conversion rate, and spending amount, as well as venue quality rating, operating hours, and promotional activities. The evaluation module is used to conduct quality access assessment and user matching assessment of the venue based on the predicted operating performance score, set grading standards according to the expected results, and filter the venues by combining the merchant site selection threshold to obtain the venue selection combination results. The monitoring module is used to monitor and analyze the quality assessment indicators, user profile matching degree and actual operating effect of the site based on the site selection and combination results, and make dynamic adjustments through early warning triggering and effect comparison to obtain the shopping mall site optimization plan.

9. A computer-readable storage medium having a computer program stored thereon, the computer program causing a processor, when executed by a processor, to perform the traffic site screening method for marketing as described in any one of claims 1 to 7.

Images