Clock sales data analysis system based on cloud

Abstract

The invention relates to the technical field of retail data analysis, in particular to a cloud-based clock sales data analysis system, which comprises a transaction time density identification module, a time sequence consumption track reconstruction module, a space hot spot dynamic detection module, an inventory fluctuation trend judgment module and an inventory linkage adjustment control module. According to the invention, through dynamic comparison of the timestamp of the clock transaction order and the number of transactions, accurate depiction of the transaction active period can be realized, and through synchronous mapping of the SKU sales number and construction of the transaction time density structure, the transaction fluctuation rule has stronger time sequence sensitivity. Through linkage analysis of geographic position codes and transaction amount change trends, formation and diffusion paths of hot spot areas are effectively captured, inventory fluctuation nodes are accurately judged in combination with inventory consumption speed and transaction amount increase trends, the inventory tension risk of the hot spot storage nodes can be effectively relieved, and the storage efficiency of the hot spot storage nodes is improved. And the space mobility and the supply and demand balance capability of the SKU inventory are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of retail data analysis, and in particular to a cloud-based watch sales data analysis system. Background Art

[0002] The field of retail data analysis technology involves the collection, organization, analysis, and mining of various types of data generated within the retail industry to support business decisions such as sales forecasting, customer behavior analysis, inventory management, and marketing optimization. This technical field primarily encompasses core issues such as data collection, data integration, data modeling and analysis, as well as visualization and report generation. By conducting in-depth analysis of multi-dimensional data such as product sales data, customer transaction records, inventory change information, and market trends, retail data analysis technology can help companies gain accurate insights into market demand, optimize operational processes, and assist in the development of precise sales and marketing strategies. Specifically, the cloud-based watch sales data analysis system relies on a cloud computing platform to centrally collect watch sales data from various sales terminals, online shopping malls, and offline stores. This data is then centrally managed through storage, and processed using rule matching and statistical report aggregation to analyze specific indicators such as sales volume, sales time, customer purchasing preferences, and best-selling product rankings.

[0003] Existing technologies rely on relational databases based on fixed structures for data storage, and use preset report templates and static query statements to perform classification statistics and time series analysis on sales data. They lack the ability to continuously and dynamically identify fluctuations in transaction time. As a result, in scenarios where sales peaks suddenly occur or demand surges in a short period of time, it is impossible to timely perceive active transaction time periods and quickly reflect the temporal changes in customer purchasing behavior. The identification of spatial hotspots relies on static geographic dimensions, making it difficult to capture the real-time diffusion trend of transaction hotspots. As a result, when inventory in hot spots is tight, effective inventory scheduling and transshipment cannot be carried out in a timely manner, affecting the dynamic balance of the overall supply chain. Summary of the Invention

[0004] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a cloud-based watch sales data analysis system.

[0005] To achieve the above objectives, the present invention adopts the following technical solutions: a cloud-based watch sales data analysis system comprising: The transaction time density recognition module is based on the time stamps and transaction counts of clock transaction orders. It divides the time stamps into time slices according to the hour of the day and the weekday classification. It extracts the transaction counts of the time slices and continuously compares them with the transaction counts of the previous time slice to generate the time density structure features of clock transactions. The time series consumption trajectory reconstruction module uses the time density structure feature of the watch transaction to extract the SKU sales quantity and customer transaction frequency. Based on the synchronous increase of the SKU sales quantity and the number of transactions within the time slice, the module defines the transaction window and generates the time trajectory sequence feature of the watch transaction. The spatial hotspot dynamic detection module uses the time trajectory sequence features of the watch transaction to extract the transaction location code and transaction amount, analyzes the trend of the geographical code transaction amount, determines the hotspot area and spatial diffusion direction, and constructs a watch sales spatial hotspot evolution model; The inventory fluctuation trend determination module calls the watch sales spatial hotspot evolution model to extract the SKU inventory quantity, inventory consumption rate and replenishment time window, and generates the watch SKU inventory fluctuation distribution status based on the synchronization relationship between the spatial hotspot transaction growth trend and the SKU inventory consumption rate.

[0006] As a further solution of the present invention, the time density structure characteristics of watch transactions include the distribution of active transaction periods, the change pattern of transaction density, the incremental degree of the number of transactions, and the synchronization trend of SKU sales; the time trajectory sequence characteristics of watch transactions include the transaction window sequence, transaction frequency pattern, SKU sales high-frequency period, and transaction active time path; the watch sales space hotspot evolution model includes the hotspot area geocoding, transaction amount diffusion trend, spatial activity intensity change, hotspot formation and decay cycle; the watch SKU inventory fluctuation distribution status includes inventory high consumption nodes, inventory tight risk points, inventory fluctuation frequency, and SKU inventory remaining estimation results.

[0007] As a further solution of the present invention, the transaction time density identification module includes: The time slice division submodule is based on the time stamp of the clock transaction order and the number of transactions. It distinguishes the transaction order timestamps according to the hours of the day and the weekdays. It divides the transaction order timestamps into multiple time slices, marks each divided time slice with a serial number, and obtains the time slice transaction number sequence; The continuous comparison and mapping submodule selects the number of transactions in each time slice and the number of transactions in the previous time slice according to the sequence of the number of transactions in the time slice, calculates the difference, and performs numerical mapping of the corresponding time slice based on the difference calculation result and the SKU sales quantity to generate the time density structure characteristics of watch transactions.

[0008] As a further solution of the present invention, the time series consumption trajectory reconstruction module includes: The sales frequency extraction submodule calls the time slice transaction count of the watch transaction time density structure feature, aggregates and counts the transaction count according to the customer's transaction record, matches the customer's transaction count with the SKU sales quantity, and generates a comparison table of customer transaction frequency and SKU sales quantity; The synchronization window determination submodule detects whether the number of transactions and the number of SKU sales are increasing at the same time in the time slice dimension based on the customer transaction frequency and SKU sales volume comparison table, demarcates and identifies the transaction window, and obtains the transaction window time slice sequence; The time trajectory submodule calls the transaction window time slice sequence, extracts the number of transactions and SKU sales within the corresponding time slice according to the customer's transaction window time slice sequence, calculates the transaction window offset impact value, identifies the customer's complete time evolution trajectory, and generates the watch transaction time trajectory sequence features.

[0009] As a further solution of the present invention, the transaction window offset impact value adopts the formula: ; in, On behalf of the client In the trading window The impact value of the transaction window offset of the time slice, On behalf of the client In the trading window The number of transactions in a time slice, On behalf of the client In the trading window Time slice The number of SKUs sold per transaction. On behalf of the client In the trading window Time slice The cumulative number of SKUs in transactions, On behalf of the client In the trading window Time slice The transaction amount of the transaction.

[0010] As a further solution of the present invention, the spatial hotspot dynamic detection module includes: The geocoding extraction submodule calls the time trajectory sequence features of the watch transaction to extract the transaction geographic location code and the corresponding transaction amount, classifies the transaction records according to the geographic location code, and obtains a geographic location transaction amount summary table; The transaction trend calculation submodule extracts the transaction amount sequence under the same geographic code according to the geographic location transaction amount summary table and the order of time slices in the time trajectory sequence, calculates the change range of the geographic code transaction amount between adjacent time slices, and obtains the geographic code transaction change trend; The hotspot diffusion determination submodule calls the geocoded transaction change trend, identifies the geocoded areas where the transaction amount continues to rise based on the positive trend of the transaction amount change in continuous time slices, identifies the dynamic evolution structure of the watch sales space hotspot in the time dimension and the space dimension, and generates a watch sales space hotspot evolution model.

[0011] As a further solution of the present invention, the variation range of the geocoded transaction amount between adjacent time slices is calculated using the formula: ; in, Represents geocoding as The region in The transaction change range of the time slice, Represents geocoding as The region in The transaction amount of the time slice, Represents geocoding as The region in The transaction amount of the time slice, Represents geocoding as The region in The transaction amount of the time slice, Representative Order weighting coefficient, Represents the number of time slices, Represents geocoding as The region in Before time slice The average transaction amount in a time slice, Represents a positive constant.

[0012] As a further solution of the present invention, the inventory fluctuation trend determination module includes: The inventory parameter extraction submodule calls the watch sales spatial hotspot evolution model and extracts the inventory quantity of the SKU under the corresponding geographic code based on the time slice and spatial hotspot information. It also extracts the inventory consumption rate and replenishment time window based on the transaction time slice. The watch inventory quantity, inventory consumption rate and replenishment time window are mapped to the geographic code according to the SKU to generate SKU inventory dynamic change information; The synchronization relationship identification submodule extracts the transaction amount growth trend of the corresponding spatial hotspot based on the dynamic change information of the SKU inventory, pairs the transaction amount growth trend with the SKU inventory consumption rate, calculates the synchronization change degree based on the synchronization of the two within the time slice, and classifies the synchronization change degree according to SKU and geographic code to obtain the transaction inventory synchronization change trend value; The fluctuation node determination submodule calls the peak fluctuation interval of the transaction inventory synchronization change trend value to identify the time node when the watch inventory consumption rate changes. At the same time, it determines whether the inventory is in an active change state based on the replenishment time window, and obtains the key nodes of SKU inventory fluctuation; The inventory distribution submodule is based on the key nodes of SKU inventory fluctuations. According to the combination of SKU and geographic code, the inventory fluctuation nodes are mapped in the time dimension and the space dimension, and the dynamic change information of inventory quantity, inventory consumption speed and replenishment time window is integrated to generate the inventory fluctuation distribution status of watch SKU.

[0013] As a further solution of the present invention, the system further includes an inventory linkage adjustment control module: The inventory linkage adjustment control module calls the watch SKU inventory fluctuation distribution status, obtains the warehouse SKU inventory quantity, inventory consumption rate and inventory transfer capacity, and judges whether the SKU has formed an inventory shortage state at the target warehouse node based on the change trend of the inventory consumption rate compared with the transfer capacity, triggers the inventory transfer instruction, and generates the watch SKU inventory linkage adjustment result; The watch SKU inventory linkage adjustment result includes the inventory transfer path, target inventory replenishment node, inventory adjustment priority, and available storage distribution status.

[0014] As a further solution of the present invention, the inventory linkage adjustment control module includes: The inventory status identification submodule calls the watch SKU inventory fluctuation distribution status, extracts the SKU inventory quantity, inventory consumption rate and inventory transfer capacity of the target storage node, aggregates the SKU inventory quantity and inventory consumption rate by storage node, and pairs them with the corresponding inventory transfer capacity to generate a storage SKU inventory status parameter table; The tight node determination submodule analyzes the changing trend between the inventory consumption rate and the inventory transfer capacity for each warehouse node based on the warehouse SKU inventory status parameter table. When the inventory consumption rate exceeds the inventory transfer capacity, it is determined to be in an inventory tight state. The non-hot warehouse nodes where the inventory consumption rate is lower than the inventory transfer capacity are screened, and the transferable SKU inventory is extracted to obtain the SKU inventory transfer matching record. The linkage adjustment output submodule combines the target storage node, the transfer-out storage node and the corresponding SKU inventory transfer quantity based on the SKU inventory transfer matching record, constructs the inventory flow path according to the transfer relationship, and generates the watch SKU inventory linkage adjustment result according to the serialized transfer instructions of the SKU and the storage node.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, by dynamically comparing the timestamp of watch transaction orders with the number of transactions, it is possible to accurately depict the active transaction period. By synchronously mapping the SKU sales quantity, a transaction time density structure is constructed, so that the transaction fluctuation pattern has a stronger time-series sensitivity. Based on the synchronous increase in transaction frequency and SKU sales quantity, the transaction window can be dynamically delineated to accurately reflect the changes in customer consumption behavior in a specific time period. Through the linkage analysis of geographic location coding and transaction amount change trends, the formation and diffusion paths of hot spots can be effectively captured. Combined with the inventory consumption rate and the transaction amount growth trend, the inventory fluctuation nodes can be accurately determined, and dynamic control of inventory transfer paths, replenishment nodes and adjustment priorities can be achieved. This can effectively alleviate the inventory shortage risk of hot storage nodes and improve the spatial liquidity and supply and demand balance capabilities of SKU inventory. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 Schematic diagram of a cloud-based watch sales data analysis system according to the present invention; Figure 2 Schematic diagram of the system framework in the present invention; Figure 3 This is a schematic diagram of the transaction time density identification module in the present invention; Figure 4 This is a schematic diagram of the time series consumption trajectory reconstruction module in the present invention; Figure 5 Schematic diagram of the spatial hotspot dynamic detection module in the present invention; Figure 6 This is a schematic diagram of the inventory fluctuation trend determination module in the present invention; Figure 7 Schematic diagram of the inventory linkage adjustment control module in the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0018] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0019] See also Figure 1 、 2 , a cloud-based watch sales data analysis system includes: The transaction time density recognition module is based on the timestamps and transaction counts of watch transaction orders. It divides the timestamps into time slices according to the hour of the day and the weekday classification. It extracts the transaction counts in the time slices and continuously compares them with the transaction counts in the previous time slice. It then synchronizes the mapping with the SKU sales quantity to generate the structural characteristics of the watch transaction time density. The time-series consumption trajectory reconstruction module uses the time density structure characteristics of watch transactions to extract SKU sales quantity and customer transaction frequency. Based on the synchronous increase in SKU sales quantity and transaction number within the time slice, it delineates the transaction window and generates the time trajectory sequence characteristics of watch transactions. The spatial hotspot dynamic detection module uses the time trajectory sequence characteristics of watch transactions to extract the transaction location code and transaction amount, analyze the changing trend of the geocoded transaction amount, determine the hotspot area and spatial diffusion direction, and build a watch sales spatial hotspot evolution model; The inventory fluctuation trend determination module uses the watch sales spatial hotspot evolution model to extract SKU inventory quantity, inventory consumption rate, and replenishment time window. Based on the synchronous relationship between the spatial hotspot transaction growth trend and SKU inventory consumption rate, it determines the active inventory fluctuation nodes and generates the watch SKU inventory fluctuation distribution status. The inventory linkage adjustment control module calls the watch SKU inventory fluctuation distribution status to obtain the warehouse SKU inventory quantity, inventory consumption rate and inventory transfer capacity. Based on the changing trend of inventory consumption rate compared with transfer capacity, it determines whether the SKU has formed an inventory shortage state at the target storage node, triggers the inventory transfer instruction, and generates the watch SKU inventory linkage adjustment results; The time density structure characteristics of watch transactions include the distribution of active transaction periods, the change pattern of transaction density, the incremental degree of transaction number, and the synchronization trend of SKU sales. The time trajectory sequence characteristics of watch transactions include the transaction window sequence, transaction frequency pattern, SKU sales high-frequency period, and transaction active time path. The watch sales spatial hotspot evolution model includes the geocoding of hotspot areas, transaction amount diffusion trend, spatial activity intensity change, hotspot formation and decay cycle. The watch SKU inventory fluctuation distribution status includes inventory high consumption nodes, inventory shortage risk points, inventory fluctuation frequency, and SKU inventory remaining estimation results. The watch SKU inventory linkage adjustment results include inventory transfer path, target inventory replenishment node, inventory adjustment priority, and available storage distribution status.

[0020] See also Figure 2 、 3 , the transaction time density identification module includes: The time slice division submodule is based on the time stamp of the clock transaction order and the number of transactions. It distinguishes the transaction order timestamps according to the hours of the day and the weekdays. It divides the transaction order timestamps into multiple time slices, marks each divided time slice with a serial number, and obtains the time slice transaction number sequence; Extract the original order data from the trading platform database. The timestamp field represents the specific time when the transaction occurred, and the transaction number field represents the number of transactions corresponding to a certain point in time. After importing the data into the data analysis platform, unify the format of the timestamp field to ensure that the time format conforms to the standard of year-month-day-hour:minute:second. Set the order time "2025-06-03-09:15:00" to clearly represent 9:15 am on June 3, 2025. Use the programming tool to extract the hour information of each order record and determine whether it corresponds to a weekday or a weekend. The judgment method is based on the weekday number. Days greater than or equal to five are marked as weekdays, and days equal to or greater than five are marked as weekends. Based on the extracted hourly time periods, the period between 9 and 10 o'clock is set as the 9th time period, and the orders are classified according to "weekday-hour period" or "weekend-hour period". After classification, the total number of transactions under each category is counted. It is assumed that there are 55 transactions under the "weekday-9th time period" and 32 transactions under the "weekend-14th time period". After classification, the time slices are numbered in sequence, for example, numbered from 1 to N according to chronological order. The numbering process is sorted from the earliest time period to the latest time period to obtain a time slice transaction number sequence.

[0021] The continuous comparison and mapping submodule calculates the difference between the number of transactions in each time slice and the number of transactions in the previous time slice based on the sequence of transaction counts in the time slice. Based on the difference calculation results and the SKU sales quantity, the numerical mapping is performed on the corresponding time slice to generate the time density structure characteristics of watch transactions. By traversing the transaction number sequence, the transaction number of the current time slice is directly subtracted from the transaction number of the previous time slice, and the difference in the change in the transaction number between each consecutive time slice is calculated. In this way, the fluctuation trend of the transaction number can be clearly captured. If the transaction number of a certain time slice is set to 50, and the transaction number of the previous time slice is 30, then the change value is 20, indicating that the transaction number has increased. If the opposite is true, the change value is negative, indicating that the transaction number has decreased. After completing the difference calculation, the SKU sales quantity information corresponding to each time slice is introduced. The SKU sales quantity is obtained through the detailed statistics of the order items. If a total of 8 SKUs are sold in a certain time slice, then the time The number of SKUs sold in a time slice is 8. Based on the combination of the trend of changes in the number of transactions and the number of SKUs sold, a mapping rule is formulated. If the change in the number of transactions is positive and the number of SKUs sold is high, a positive numerical mapping is set. If the change in the number of transactions is negative and the number of SKUs sold is low, a negative numerical mapping is set. For other cases, a neutral value is assigned according to the actual situation. Each time slice is mapped and marked according to the rules. After the mapping is completed, the mapped value sequence is windowed and slidingly combined. A combination is formed every three consecutive time slices, and the combination is slid in sequence until the entire sequence is traversed. This combination can reflect the changing pattern of transaction behavior under time continuity and generate the time density structure characteristics of clock transactions.

[0022] See also Figure 2 、 4 ,The time series consumption trajectory reconstruction module includes: The sales frequency extraction submodule uses the time-slice transaction counts of the clock transaction time density structure feature, aggregates and counts the transaction counts according to the customer's transaction records, and matches the customer's transaction counts with the SKU sales quantity to generate a comparison table of customer transaction frequency and SKU sales quantity; Based on the customer's unique identifier, the transaction records of the customer in each time slice are extracted from the order database. The two core fields of the time slice transaction count and the corresponding SKU sales quantity are read. The transaction count is grouped and counted according to the customer dimension to obtain the total number of transactions for each customer in each time slice. The transaction count of customer A in the "weekday-9 period" is set to 5, and the transaction count in the "weekday-14 period" is set to 2. The time slices are traversed in turn. At the same time, the SKU sales quantity is obtained by accumulating the number of SKUs purchased by the customer in the corresponding time slice. For example, in the "weekday-9 period", customer A purchased 3 different SKUs. The cumulative sales volume of KU's products is 6. After completing the two-way aggregation statistics of the number of transactions and the number of SKU sales, the two are matched one by one according to the time slice. The comparison table uses time slices as rows and the number of transactions and the number of SKU sales as columns. It is set as follows: time slice "weekday-9 period", number of transactions 5, number of SKU sales 6; time slice "weekday-14 period", number of transactions 2, number of SKU sales 3. Through such a comparison table, the relationship between the customer's transaction activity and commodity purchase intensity in different time slices can be fully described, providing a standardized input basis for subsequent transaction pattern recognition, and generating a comparison table of customer transaction frequency and SKU sales volume.

[0023] The synchronization window determination submodule detects whether the number of transactions and SKU sales are increasing simultaneously in the time slice dimension based on the comparison table of customer transaction frequency and SKU sales volume. It then delineates and identifies the transaction window and obtains the transaction window time slice sequence. The number of transactions and SKU sales in each time slice are tested one by one. The judgment standard is based on the change trend of two adjacent time slices in the time series. The specific operation is to compare the number of transactions in the current time slice with the previous time slice, and at the same time compare the number of SKU sales in the current time slice with the previous time slice. When both show an upward trend, it is determined that the time slice is in the transaction synchronization rising window. Suppose a customer has 5 transactions and 6 SKU sales in the time slice "weekday-9 period", and the number of transactions rises to 7 and the number of SKU sales rises to 9 in the "weekday-10 period". , then the period from 9 to 10 is judged to belong to the synchronous transaction rising window. If the number of transactions increases but the SKU sales quantity decreases, or vice versa, it is not determined to be a synchronous window. After the judgment is completed, the time slices that meet the conditions are classified and marked. The time slice sequence is arranged in chronological order, such as "weekday-9 period", "weekday-10 period", and "weekday-11 period". By detecting the continuity of the sequence, continuous transaction peak periods can also be identified and classified in the form of window labels. For example, if three consecutive time slices rise synchronously, they are marked as high-frequency trading windows, and the trading window time slice sequence is output.

[0024] The time trajectory submodule calls the transaction window time slice sequence, extracts the number of transactions and SKU sales within the corresponding time slice according to the customer's transaction window time slice sequence, calculates the transaction window offset impact value, identifies the customer's complete time evolution trajectory, and generates the watch transaction time trajectory sequence features; The impact value of trading window offset is calculated using the formula: ; in, On behalf of the client In the trading window The impact value of the transaction window offset of the time slice, On behalf of the client In the trading window The number of transactions in a time slice, On behalf of the client In the trading window Time slice The number of SKUs sold per transaction. On behalf of the client In the trading window Time slice The cumulative number of SKUs in transactions, On behalf of the client In the trading window Time slice The transaction amount of the transaction; Formula calculation logic: By integrating the SKU sales quantity of each transaction with the cumulative occurrence frequency of the SKU in the current time slice, a weighted sales impact term is constructed. The difference between this term and the total actual transaction amount is then calculated. The difference is then normalized by division after adding 1 to the number of transactions, and the absolute value is taken. The number of SKUs is used to measure transaction intensity, the square root of the SKU frequency reflects the nonlinear impact of sales hotspots, and the transaction amount reflects the customer's actual consumption behavior. By normalizing the difference between the weighted SKU structure and the amount, the degree of structural deviation between SKU sales behavior and payment amount in the time slice is revealed. The entire calculation logic focuses on the degree of mismatch between the sales structure formed by multiple transactions and the value input within the transaction time slice, which is used to capture the non-equilibrium characteristics of consumption behavior and provide a structural quantitative basis for the subsequent construction of time evolution trajectory. The transaction window deviation impact value measures the average difference between the weighted SKU structure and the actual transaction amount of a customer within a certain transaction time slice. This indicator reflects the structural deviation between SKU sales activity and capital payment within that time slice. Parameter meaning and formula calculation derivation process: Customer transaction data is collected through the linkage between retail and front-end terminals and continuously monitored by time slice. Items collected include the number of SKUs per transaction, the number of times the SKU appears in the time slice, and the amount of each transaction. Take the third time slice of the second trading window for example: Parameter acquisition process and quantification standards: :The front-end transaction system has a built-in log capture function, and the transaction count counter in each time slice is recorded. The third time slice records a total of 4 transactions, so ; : The SKU sales quantity is extracted from the sales records returned by the product barcode scanner in the sales system. The SKU sales quantity is defined as the actual settlement quantity in the transaction. The transaction records are as follows: 3 for the first transaction, 2 for the second transaction, 5 for the third transaction, and 4 for the fourth transaction. The cumulative number of SKU occurrences is tracked using a real-time SKU frequency tracking device. The cumulative frequency of each SKU in each transaction during the time slice is calculated as follows: 9 for the first transaction, 4 for the second, 16 for the third, and 1 for the fourth. : The transaction amounts come from the POS system and are settled synchronously. The first transaction is 120 yuan, the second is 86 yuan, the third is 210 yuan, and the fourth is 92 yuan. Parameter substitution: ; ; Substitute into the main formula: ; The results show that in the third time slice of the second transaction window, the transaction window offset impact value between the SKU quantity, SKU popularity value and transaction amount of customer A is 91.4, reflecting that there is a strong difference in the weighted SKU structure of the transaction amount. The larger the value, the greater the mismatch between the SKU structure in the time slice and the actual transaction amount. In the subsequent time trajectory sequence, this value will participate in the construction of the transaction offset trajectory feature dimension to support the structural composition of the evolution trajectory.

[0025] See also Figure 2 、 5 ,The spatial hotspot dynamic detection module includes: The geocoding extraction submodule uses the time trajectory sequence features of clock transactions to extract the transaction location codes and corresponding transaction amounts. It then categorizes the transaction records according to the location codes to generate a summary table of location transaction amounts. The geographic location code is extracted based on the geographic location information contained in the transaction record. The code is a raster code converted from the transaction store number, postal code, or GPS longitude and latitude. If the transaction occurs in Jing'an District, Shanghai, it can be converted to "310106" through geographic coding. At the same time, the corresponding transaction amount is extracted from the transaction record. The transaction amount field directly corresponds to the transaction value of each transaction. The transactions are classified according to the geographic code, and the transaction records in different time periods under the same geographic code are summarized and counted. Set within the "working day-9 period", there are three transactions under the geographic code "310106", with transaction amounts of 1,200 yuan, 800 yuan, and 1,500 yuan respectively, and the total amount is 3,500 yuan. The transactions in the time slice are classified and summarized in the same way to form a geographic location transaction amount summary table.

[0026] The transaction trend calculation submodule extracts the transaction amount sequence under the same geographic code according to the order of time slices in the time trajectory sequence based on the geographic transaction amount summary table, calculates the change range of the geocoded transaction amount between adjacent time slices, and obtains the geocoded transaction change trend; The change in geocoded transaction amounts between adjacent time slices is calculated using the formula: ; in, Represents geocoding as The region in The transaction change range of the time slice, Represents geocoding as The region in The transaction amount of the time slice, Represents geocoding as The region in The transaction amount of the time slice, Represents geocoding as The region in The transaction amount of the time slice, Representative Order weighting coefficient, Represents the number of time slices, Represents geocoding as The region in Before time slice The average transaction amount in a time slice, Represents a positive constant; The calculation logic is as follows: The transaction change amplitude is based on the difference between the current and previous time slice transaction amounts, superimposed on the weighted sum of the differences between the transaction amounts in several time slices and their sliding averages, to reflect the combined effects of short-term fluctuations and deviations. The overall result is then normalized by adding a very small square root to the absolute value of the previous time slice transaction amount to compress differences of different orders of magnitude and output as an absolute value to ensure a magnitude expression without directional deviation. Transaction change amplitude refers to the intensity of changes in transaction amounts within the same geographic region over consecutive time slices, measuring the dynamic volatility of transaction behavior in that region. It is obtained by comparing the current transaction amount with the amount in the previous time slice and incorporating the weighted deviation of the time slice, after normalization. Parameter meaning and formula calculation derivation process: Transaction amount data is collected through city-level POS terminal transaction records, with the unit being 10,000 yuan. The time slice span is 24 hours, and the sampling time point is set according to the sampling frequency of once every 24 hours. The geographic code is set to a=0755 to represent Shenzhen City, and the time slice t=5 is selected. The historical data of K=3 time slices are taken forward. The actual sampling data is as follows: Ten thousand yuan (transaction amount in the 5th time slice), Ten thousand yuan, Ten thousand yuan, Ten thousand yuan, Ten thousand yuan; Weighting coefficient The time decay degree is set to: ,because The most recent transaction has the greatest impact and is set as the benchmark weight; , the influence is second strongest due to time lag; ,Due to decreasing impact, the lowest impact is given the lowest weight; Average transaction amount 、 、 The values ​​are calculated by taking the weighted average of the transaction data of the three consecutive days, as follows: Ten thousand yuan; Ten thousand yuan; Ten thousand yuan; In the denominator To avoid the denominator being 0, the minimum positive value standard is set as: ; Substitute the formula and expand the calculation as follows: 1. Calculate the differences and weighted terms: ; , multiplied by get: ; , multiplied by Result: 3.771; , multiplied by Result: 0.074; Add up the numerator: molecular: ; Denominator calculation: Denominator: ; Evaluate the overall expression: ; The result shows that the transaction change amplitude of the area with the geographic code of 0755 in time slice 5 is 1.76, which reflects the degree of the amount change trend based on the previous day and is used as a trend factor to participate in the next step of transaction trend extraction.

[0027] The hotspot diffusion determination submodule uses geocoded transaction change trends to identify geocoded areas with continuously rising transaction amounts based on the positive trend of transaction amounts in consecutive time slices. It also identifies the dynamic evolution structure of watch sales hotspots in the temporal and spatial dimensions and obtains a watch sales hotspot evolution model. The diffusion paths of each transaction hotspot geocoded are serialized according to the order of time slices. The hotspot location, transaction amount change trend, and diffusion direction within each time slice are integrated to form a multi-dimensional evolution sequence. For example, the hotspot location is set at "310106" during the "weekday-9 period" and the diffusion direction is northward. The transaction amount continues to rise, spreading to "310107" during the "weekday-10 period" and to "310108" during the "weekday-11 period". By connecting the hotspot changes within each time slice through the time axis, a dynamic spatial diffusion path diagram can be constructed. This path not only reflects the spatial movement of the hotspot but also the speed and direction of change in the transaction amount. By analyzing the sequence, it can be found that some hotspots have a rapid diffusion trend, while others show a relatively slow or even stagnant state. In the spatial evolution model, each node represents a specific geocoded hotspot, and each edge represents the diffusion connection between hotspots in adjacent time slices. This can fully describe the coordinated changes in transaction hotspots in the temporal and spatial dimensions, providing accurate data support for sales forecasting, market analysis, and marketing strategy optimization, and generating a spatial hotspot evolution model for watch sales.

[0028] See also Figure 2 、 6 ,The inventory fluctuation trend determination module includes: The inventory parameter extraction submodule uses the watch sales spatial hotspot evolution model to extract the inventory quantity of the SKU under the corresponding geocode based on time slice and spatial hotspot information. It then extracts the inventory consumption rate and replenishment time window based on the transaction time slice. The watch inventory quantity, inventory consumption rate, and replenishment time window are mapped to the geocode according to the SKU to generate SKU inventory dynamic change information. Based on the combination of each time slice and the spatial hotspot geocoding in the hotspot evolution model, the inventory quantity of the corresponding SKU is extracted. The inventory data comes from the real-time inventory database of the warehouse management equipment or the store. Under the geocoding "310106", the inventory quantity of SKU A in the "working day-9 period" is 120. Through the transaction records of consecutive time slices, the inventory consumption rate of the SKU under the geocoding can be analyzed. For example, 10 units are sold in the 9 period, and the inventory drops to 110 in the 10 period. In this way, the inventory reduction in each time slice is calculated, and the replenishment time window is identified in combination with the replenishment record. It is set in the "working day-11 period" when the inventory is detected to be reduced from 100 to 110. If the number of items increases to 200, it means that replenishment occurs within this time slice. The inventory quantity, inventory consumption rate and replenishment time window are matched, and then combined according to the SKU and geographic code. The inventory starting value, consumption rate and replenishment time point of the SKU in different geographic locations within each time slice are recorded. For example, SKU-A with the geographic code "310106" has 120 items in inventory in time period 9, a consumption rate of 10, and an empty replenishment time window; in time period 11, the inventory increases from 100 to 200, and the replenishment time window is marked as the current time period. Through continuous tracking, the dynamic changes of SKU inventory in spatial hotspot areas can be fully reflected, and SKU inventory dynamic change information can be generated.

[0029] The synchronization relationship identification submodule extracts the transaction amount growth trend of the corresponding spatial hotspot based on the dynamic change information of SKU inventory. It then pairs the transaction amount growth trend with the SKU inventory consumption rate. Based on the synchronization of the two changes within the time slice, it calculates the synchronization change degree. The synchronization change degree is classified by SKU and geocode to obtain the transaction inventory synchronization change trend value. According to the same combination of geocoding and time slices, the direction of change of the two variables is judged. If the transaction amount shows an upward trend in the continuous time slices and the consumption rate of SKU inventory is also accelerating, it is determined to be a synchronous positive change. If the transaction amount decreases and the inventory consumption slows down, it is also considered a synchronous change. On the contrary, if the transaction amount increases but the inventory consumption rate decreases, or the transaction amount decreases and the inventory consumption accelerates, it is considered an asynchronous change. By traversing the SKU combination under the time slice and geocoding, the synchronous change degree of each SKU in the spatial hotspot area is calculated, and the synchronous change degree is expressed as Numerical classification is performed under the combination of SKU and geocode to form the transaction-inventory synchronization change trend value. Set SKU-A under the geocode "310106", the transaction amount increases by 15% from period 9 to period 10, and the inventory consumption rate increases from 10 per hour to 15 per hour. The synchronization change degree is marked as high. If the transaction amount increase drops to 5% in the time slice, but the inventory consumption rate remains at 15, the synchronization change degree drops to medium. Through dynamic matching, the linkage between transactions and inventory can be effectively captured, and the transaction-inventory synchronization change trend value can be obtained.

[0030] The fluctuation node determination submodule calls the peak fluctuation interval of the transaction inventory synchronization change trend value to identify the time node when the watch inventory consumption rate changes. At the same time, it combines the replenishment time window to determine whether the inventory is in an active change state and obtain the key nodes of SKU inventory fluctuation; A time series analysis is performed on the synchronous change trend values ​​of each SKU and geocode combination to identify peaks, valleys, and fluctuation mutation points. These are set in the time series "high, medium, high, low, high." Switching points from high to low and then back to high are identified as fluctuation nodes. Cross-validation is performed based on the replenishment time window. If the fluctuation node happens to be in the time period before and after replenishment and the inventory consumption rate changes significantly, it is determined to be a valid fluctuation node. For example, for SKU-A with the geocode "310106", the inventory consumption rate suddenly increases from 12 to 20 per hour during the "weekday-10 period". At the same time, the transaction amount growth trend also fluctuates significantly. However, in the "weekday-11 period", the consumption rate drops to 8, accompanied by inventory replenishment. This sudden increase followed by a sudden drop is clearly identified as a key inventory fluctuation node. By analyzing the fluctuation intervals in the time series one by one and combining inventory changes, transaction trends, and replenishment windows, we can accurately locate the time nodes when inventory changes are most drastic and identify the key nodes of SKU inventory fluctuation.

[0031] The inventory distribution submodule maps the inventory fluctuation nodes in time and space based on the combination of SKU and geocoding based on the key nodes of SKU inventory fluctuation. It integrates the dynamic change information of inventory quantity, inventory consumption speed and replenishment time window to generate the inventory fluctuation distribution status of watch SKU. According to the combination of SKU and geocoding, the fluctuation nodes are visually mapped in the time dimension and space dimension, and the time point of inventory fluctuation is corresponded to the spatial location. Through this mapping, the distribution status of inventory fluctuations in the entire sales network can be intuitively displayed. At the same time, the dynamic change information of inventory quantity, inventory consumption rate and replenishment time window is integrated to establish a complete inventory change trajectory model for each SKU under different geocoding. It is set that SKU-A in geocoding "310106" has an inventory of 120 units in "weekday-9 period", which drops to 100 units in "10 period" and is replenished to 200 units in "11 period". In the same time period, the inventory of geocoding "310107" first decreases slowly and then there is a short replenishment. The asynchrony of inventory fluctuations in different spatial locations is fully demonstrated, generating the inventory fluctuation distribution status of watch SKUs.

[0032] See also Figure 2 、 7 , the inventory linkage adjustment control module includes: The inventory status identification submodule uses the watch SKU inventory fluctuation distribution status to extract the target storage node SKU inventory quantity, inventory consumption rate and inventory transfer capacity. The SKU inventory quantity and inventory consumption rate are aggregated by storage node, and paired with the corresponding inventory transfer capacity to generate a storage SKU inventory status parameter table; Based on the inventory quantity, inventory consumption rate, and inventory transfer capacity of the SKU at different storage nodes recorded in the inventory distribution status, detailed inventory information of the current target storage node is extracted. The inventory quantity is directly derived from the warehouse's real-time inventory data. The inventory consumption rate is obtained by counting the inventory reduction and transaction activity within the time slice. The inventory transfer capacity is extracted based on the logistics processing capacity, loading and unloading capacity, and vehicle allocation of the storage node. Assume that the current inventory quantity of SKU A at a certain storage node "WH001" is 500, the inventory consumption rate is 80 per day, and the inventory transfer capacity is 100 per day. The inventory quantity and consumption rate of SKUs at the same storage node are aggregated and paired with the aggregated SKU inventory data of the storage node's transfer capacity. If the inventory consumption rate of a SKU is far lower than the transfer capacity, the SKU is considered to be in a safe stock state. Conversely, if the inventory consumption is close to the transfer upper limit, it is marked as a high-risk state, and a storage SKU inventory status parameter table is generated.

[0033] The tight node determination submodule analyzes the changing trend between inventory consumption rate and inventory transfer capacity for each warehouse node based on the warehouse SKU inventory status parameter table. When the inventory consumption rate exceeds the inventory transfer capacity, it is determined to be in an inventory tight state. Non-hot warehouse nodes with inventory consumption rates lower than the inventory transfer capacity are screened, and the transferable SKU inventory is extracted to obtain SKU inventory transfer matching records. For each tight storage node, traverse the non-hot storage nodes and select the storage nodes with inventory consumption rate lower than the inventory transfer capacity as potential transfer-out nodes. Set the current SKU-A inventory of storage node "WH002" to 800, the inventory consumption rate is only 50 per day, and the transfer capacity is 120 per day. There is obviously surplus inventory and sufficient transfer capacity. Pair the transferable inventory of this node with the target tight node to determine whether the target storage node has the inventory receiving capacity. The judgment standard is whether the maximum storage capacity of the node minus the current inventory quantity is greater than the number of SKUs to be transferred in. If the conditions are met, directly establish a transfer matching relationship. Set the current inventory of target node "WH001" to 500, the maximum capacity to 1000, and the number of SKU-A to be transferred in to 200, which meets the acceptance conditions. If not, continue to traverse the next non-hot node "WH003" and repeat the above judgment process until the matching task of the entire SKU inventory transfer is completed and the SKU inventory transfer matching record is generated.

[0034] The linkage adjustment output submodule combines the target storage node, the transfer-out storage node, and the corresponding SKU inventory transfer quantity based on the SKU inventory transfer matching record, constructs the inventory flow path based on the transfer relationship, and generates the watch SKU inventory linkage adjustment results according to the serialized transfer instructions of the SKU and the storage node; According to each matching relationship, the target storage node, the transfer storage node and the corresponding SKU inventory transfer quantity are combined, and the inventory flow path is automatically generated according to the transfer path logic. The path planning not only considers the geographical distance, but also the transportation time, traffic conditions and the loading and unloading efficiency of the transfer node. SKU-A is transferred from the storage node "WH002", the target storage node is "WH001", and the path is "WH002-city highway-WH001". At the same time, specific transfer batches are arranged according to the current availability and distribution capacity of the transfer vehicles. According to the transportation time window, the transfer instructions are serialized according to the combination order of SKU and storage node to form a structured inventory linkage adjustment instruction. The instruction content is set to include "allocate 200 SKU-A quantities from WH002 to WH001, the estimated outbound time is 10:00, and the estimated inbound time is 14:00", which reflects the dynamic status of the current inventory adjustment network, ensures that the inventory can be efficiently circulated in the entire supply chain network, guarantees the supply capacity of hot spots, alleviates the inventory pressure of tense nodes, and generates the watch SKU inventory linkage adjustment results.

[0035] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A cloud-based watch sales data analysis system, characterized in that: The system comprises: The transaction time density recognition module is based on the time stamps and transaction counts of clock transaction orders. It divides the time stamps into time slices according to the hour of the day and the weekday classification. It extracts the transaction counts of the time slices and continuously compares them with the transaction counts of the previous time slice to generate the time density structure features of clock transactions. The time series consumption trajectory reconstruction module uses the time density structure feature of the watch transaction to extract the SKU sales quantity and customer transaction frequency. Based on the synchronous increase of the SKU sales quantity and the number of transactions within the time slice, the module defines the transaction window and generates the time trajectory sequence feature of the watch transaction. The spatial hotspot dynamic detection module uses the time trajectory sequence features of the watch transaction to extract the transaction location code and transaction amount, analyzes the trend of the geographical code transaction amount, determines the hotspot area and spatial diffusion direction, and constructs a watch sales spatial hotspot evolution model; The inventory fluctuation trend determination module calls the watch sales spatial hotspot evolution model to extract the SKU inventory quantity, inventory consumption rate and replenishment time window, and generates the watch SKU inventory fluctuation distribution status based on the synchronization relationship between the spatial hotspot transaction growth trend and the SKU inventory consumption rate.

2. The cloud-based watch sales data analysis system according to claim 1, characterized in that: The time density structure characteristics of watch transactions include the distribution of active transaction periods, the change pattern of transaction density, the incremental increase in the number of transactions, and the synchronization trend of SKU sales. The time trajectory sequence characteristics of watch transactions include the transaction window sequence, transaction frequency pattern, SKU sales high-frequency period, and transaction active time path. The watch sales spatial hotspot evolution model includes the hotspot area geocoding, transaction amount diffusion trend, spatial activity intensity change, hotspot formation and decay cycle. The watch SKU inventory fluctuation distribution status includes inventory high consumption nodes, inventory shortage risk points, inventory fluctuation frequency, and SKU inventory remaining estimation results.

3. The cloud-based watch sales data analysis system according to claim 1, characterized in that: The transaction time density identification module includes: The time slice division submodule is based on the time stamp of the clock transaction order and the number of transactions. It distinguishes the transaction order timestamps according to the hours of the day and the weekdays. It divides the transaction order timestamps into multiple time slices, marks each divided time slice with a serial number, and obtains the time slice transaction number sequence; The continuous comparison and mapping submodule selects the number of transactions in each time slice and the number of transactions in the previous time slice according to the sequence of the number of transactions in the time slice, calculates the difference, and performs numerical mapping of the corresponding time slice based on the difference calculation result and the SKU sales quantity to generate the time density structure characteristics of watch transactions.

4. The cloud-based watch sales data analysis system according to claim 3, characterized in that: The time series consumption trajectory reconstruction module includes: The sales frequency extraction submodule calls the time slice transaction count of the watch transaction time density structure feature, aggregates and counts the transaction count according to the customer's transaction record, matches the customer's transaction count with the SKU sales quantity, and generates a comparison table of customer transaction frequency and SKU sales quantity; The synchronization window determination submodule detects whether the number of transactions and the number of SKU sales are increasing at the same time in the time slice dimension based on the customer transaction frequency and SKU sales volume comparison table, demarcates and identifies the transaction window, and obtains the transaction window time slice sequence; The time trajectory submodule calls the transaction window time slice sequence, extracts the number of transactions and SKU sales within the corresponding time slice according to the customer's transaction window time slice sequence, calculates the transaction window offset impact value, identifies the customer's complete time evolution trajectory, and generates the watch transaction time trajectory sequence features.

5. The cloud-based watch sales data analysis system according to claim 4, characterized in that: The transaction window offset impact value is calculated using the formula: ； in, On behalf of the client In the trading window The impact value of the transaction window offset of the time slice, On behalf of the client In the trading window The number of transactions in a time slice, On behalf of the client In the trading window Time slice The number of SKUs sold per transaction. On behalf of the client In the trading window Time slice The cumulative number of SKUs in transactions, On behalf of the client In the trading window Time slice The transaction amount of the transaction.

6. The cloud-based watch sales data analysis system according to claim 4, characterized in that: The spatial hotspot dynamic detection module includes: The geocoding extraction submodule calls the time trajectory sequence features of the watch transaction to extract the transaction geographic location code and the corresponding transaction amount, classifies the transaction records according to the geographic location code, and obtains a geographic location transaction amount summary table; The transaction trend calculation submodule extracts the transaction amount sequence under the same geographic code according to the geographic location transaction amount summary table and the order of time slices in the time trajectory sequence, calculates the change range of the geographic code transaction amount between adjacent time slices, and obtains the geographic code transaction change trend; The hotspot diffusion determination submodule calls the geocoded transaction change trend, identifies the geocoded areas where the transaction amount continues to rise based on the positive trend of the transaction amount change in continuous time slices, identifies the dynamic evolution structure of the watch sales space hotspot in the time dimension and the space dimension, and generates a watch sales space hotspot evolution model.

7. The cloud-based watch sales data analysis system according to claim 6, characterized in that: The change in the geocoded transaction amount between adjacent time slices is calculated using the formula: ； in, Represents geocoding as The region in The transaction change range of the time slice, Represents geocoding as The region in The transaction amount of the time slice, Represents geocoding as The region in The transaction amount of the time slice, Represents geocoding as The region in The transaction amount of the time slice, Representative Order weighting coefficient, Represents the number of time slices, Represents geocoding as The region in Before time slice The average transaction amount in a time slice, Represents a positive constant.

8. The cloud-based watch sales data analysis system according to claim 6, characterized in that: The inventory fluctuation trend determination module includes: The inventory parameter extraction submodule calls the watch sales spatial hotspot evolution model and extracts the inventory quantity of the SKU under the corresponding geographic code based on the time slice and spatial hotspot information. It also extracts the inventory consumption rate and replenishment time window based on the transaction time slice. The watch inventory quantity, inventory consumption rate and replenishment time window are mapped to the geographic code according to the SKU to generate SKU inventory dynamic change information; The synchronization relationship identification submodule extracts the transaction amount growth trend of the corresponding spatial hotspot based on the dynamic change information of the SKU inventory, pairs the transaction amount growth trend with the SKU inventory consumption rate, calculates the synchronization change degree based on the synchronization of the two within the time slice, and classifies the synchronization change degree according to SKU and geographic code to obtain the transaction inventory synchronization change trend value; The fluctuation node determination submodule calls the peak fluctuation interval of the transaction inventory synchronization change trend value to identify the time node when the watch inventory consumption rate changes. At the same time, it determines whether the inventory is in an active change state based on the replenishment time window, and obtains the key nodes of SKU inventory fluctuation; The inventory distribution submodule is based on the key nodes of SKU inventory fluctuations. According to the combination of SKU and geographic code, the inventory fluctuation nodes are mapped in the time dimension and the space dimension, and the dynamic change information of inventory quantity, inventory consumption speed and replenishment time window is integrated to generate the inventory fluctuation distribution status of watch SKU.

9. The cloud-based watch sales data analysis system according to claim 1, characterized in that: The system also includes an inventory linkage adjustment control module: The inventory linkage adjustment control module calls the watch SKU inventory fluctuation distribution status, obtains the warehouse SKU inventory quantity, inventory consumption rate and inventory transfer capacity, and judges whether the SKU has formed an inventory shortage state at the target warehouse node based on the change trend of the inventory consumption rate compared with the transfer capacity, triggers the inventory transfer instruction, and generates the watch SKU inventory linkage adjustment result; The watch SKU inventory linkage adjustment result includes the inventory transfer path, target inventory replenishment node, inventory adjustment priority, and available storage distribution status.

10. The cloud-based watch sales data analysis system according to claim 9, characterized in that: The inventory linkage adjustment control module includes: The inventory status identification submodule calls the watch SKU inventory fluctuation distribution status, extracts the SKU inventory quantity, inventory consumption rate and inventory transfer capacity of the target storage node, aggregates the SKU inventory quantity and inventory consumption rate by storage node, and pairs them with the corresponding inventory transfer capacity to generate a storage SKU inventory status parameter table; The tight node determination submodule analyzes the changing trend between the inventory consumption rate and the inventory transfer capacity for each warehouse node based on the warehouse SKU inventory status parameter table. When the inventory consumption rate exceeds the inventory transfer capacity, it is determined to be in an inventory tight state. The non-hot warehouse nodes where the inventory consumption rate is lower than the inventory transfer capacity are screened, and the transferable SKU inventory is extracted to obtain the SKU inventory transfer matching record. The linkage adjustment output submodule combines the target storage node, the transfer-out storage node and the corresponding SKU inventory transfer quantity based on the SKU inventory transfer matching record, constructs the inventory flow path according to the transfer relationship, and generates the watch SKU inventory linkage adjustment result according to the serialized transfer instructions of the SKU and the storage node.

Images