Supply chain management platform based on big data

Abstract

The invention, which relates to the technical field of supply chain management, discloses a big-data-based supply chain management platform comprising a data acquisition module, a data cleaning and preprocessing module, a prediction analysis module, a supplier management module, an inventory management module, a logistics and distribution management module and a visual support module. The data acquisition module acquires key data from each platform, acquires the data in batches at regular time through an API interface and a data crawler mode and transmits the data to the data cleaning and preprocessing module; according to the invention, through multi-dimensional demand analysis of the prediction analysis module, demand fluctuation of different categories, regions and sales channels is predicted in combination with historical sales data and market trends, based on the prediction, the inventory management module can optimize the inventory level and avoid stockout or excess inventory, and through a replenishment strategy and a sales promotion means of unsalable commodities, the sales promotion efficiency is improved. The inventory pressure is reduced, and the inventory turnover rate is improved.

Description

Technical Field

[0001] This invention relates to the field of supply chain management technology, specifically to a supply chain management platform based on big data. Background Technology

[0002] Supply chain management refers to the coordination and optimization management of enterprises across various stages such as raw material procurement, production, warehousing, transportation, and distribution, in order to maximize customer satisfaction while reducing costs and improving operational efficiency. The background technologies of supply chain management platforms mainly stem from the application of technologies such as big data, cloud computing, the Internet of Things, artificial intelligence, and blockchain. Artificial intelligence technology has demonstrated powerful capabilities in areas such as demand forecasting, production scheduling, and logistics optimization. AI uses machine learning algorithms to analyze historical sales data and market trends to predict future demand changes.

[0003] With the diversification and rapid changes in market demand, traditional supply chain management faces numerous challenges, including inadequate inventory management, production planning delays, transportation delays, and supplier defaults.

[0004] To address the aforementioned technical shortcomings, a solution is proposed. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a supply chain management platform based on big data.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a big data-based supply chain management platform, comprising a data acquisition module, a data cleaning and preprocessing module, a predictive analysis module, a supplier management module, an inventory management module, a logistics and distribution management module, and a visualization support module;

[0007] The data acquisition module collects key data from various platforms. It uses API interfaces and data crawlers to collect data in batches on a regular basis and transmits it to the data cleaning and preprocessing module.

[0008] The data cleaning and preprocessing module includes a data cleaning unit and a preprocessing unit. The data cleaning unit identifies and deletes duplicate records using the SHA-256 hash algorithm. The preprocessing unit fills in missing parts of numerical data with the mean and processes categorical data through label encoding. The processed data is then transmitted to the predictive analysis module.

[0009] The predictive analytics module uses the processed data to build predictive models and uses LSTM models to capture trends, seasonality, and periodicity in long-term time series. This module conducts multi-dimensional demand analysis of categories, regions, and channels through historical sales data, market trends, and customer feedback. It also forecasts demand by analyzing seasonal changes through time series analysis, uses geographic weighted regression for regional demand forecasting, and uses cross-validation to forecast channel-level demand.

[0010] The supplier management module provides a supplier communication window, allowing suppliers to synchronize production progress and delivery information, and provides a shared file and document management window. It also adjusts supplier ratings through a weighted algorithm and displays the ratings in the supplier communication window.

[0011] The inventory management module calculates the demand for the next cycle based on the forecasting model of the forecasting analysis module and generates a procurement plan. Based on the forecasted demand data, the inventory management module adjusts the inventory level and executes a replenishment strategy when the inventory is below the safety stock. When goods are not selling well, the module analyzes the turnover cycle of the goods and uses price optimization algorithms to carry out discount promotions and cross-regional allocation to reduce inventory pressure.

[0012] The logistics and delivery management module tracks the location, speed, and estimated arrival time of delivery vehicles in real time through on-board positioning devices, and provides delivery time forecasts and vehicle location information. It also optimizes delivery cycles and delivery plans through historical delivery data and regression analysis.

[0013] The visualization support module displays key performance indicator data through dynamic charts, helping managers understand the operational status.

[0014] The data acquisition module is responsible for collecting data from various sources, including: sales volume, sales time, and sales channels from POS, CRM, and e-commerce platforms; production plans, actual data, equipment status, and raw material consumption from MES and ERP systems; transportation status, transportation time, and transportation costs from TMS logistics systems; delivery status and contract fulfillment from suppliers on supply chain management platforms; and market data from social media and industry reports, including price fluctuations and consumer trends. The module uses API interfaces and web crawlers to periodically collect data from multiple platforms in batches, transmitting the collected data to the data cleaning and preprocessing module.

[0015] The data cleaning and preprocessing module improves the quality and usability of the collected data. It includes a data cleaning unit and a preprocessing unit. The data cleaning unit identifies and deletes duplicate records using the SHA-256 hash algorithm. The preprocessing unit fills in missing parts of numerical data with the mean, then uses the Z-Score method to identify and delete outliers in the data. The commonly used threshold is set to ±3 standard deviations. Finally, Min-Max standardization is used to map the data to the [0,1] interval, and LabelEncoding is used to process categorical data, unifying the data format and standardizing the data. The processed data of each category is then transmitted to the predictive analysis module.

[0016] The predictive analysis module uses the time series analysis methods ARIMA and ExponentialSmoothing to process the data. ARIMA is used to capture short-term trends and seasonal fluctuations, while ExponentialSmoothing is used to smooth the data and predict future demand fluctuations. The LSTM model is used to model the trends, seasonality and periodicity in long-term time series, which is suitable for demand data with complex time dependencies.

[0017] The predictive analysis module performs multi-dimensional demand analysis based on historical sales data, market trends, and customer feedback, considering product category, region, and channel. It also employs time series analysis to decompose the seasonal variations of each product category. For regional demand forecasting, it uses Geographically Weighted Regression (GWR) to analyze demand characteristics in different regions and combines these with regional consumption characteristics and economic conditions to create customized forecasts. Furthermore, the predictive analysis module analyzes demand fluctuations across different sales channels and uses cross-validation to predict channel-level demand.

[0018] In the data processing, the ARIMA model is first used to differencing the historical data to eliminate trend and seasonal components. Then, Exponential Smoothing is used to smooth the stationary data in order to capture future demand fluctuation trends. The LSTM model is trained on historical demand data to capture long-term time dependencies, thereby predicting demand fluctuations and cycles and providing accurate demand forecast results.

[0019] The supplier management module is responsible for managing suppliers and improving supply chain efficiency. It provides a supplier communication window to synchronize production progress and delivery information, and a shared file and document management window to manage contracts, orders, quality reports, and documents. It tracks key performance indicators of suppliers, such as delivery time, product quality, and customer service level. Based on historical data, it adjusts supplier scores through a weighted algorithm and displays the supplier scores in the supplier communication window.

[0020] The inventory management module calculates the demand for the next cycle based on the predictive model in the predictive analysis module and generates a procurement plan. It adjusts the inventory according to the predicted demand data. When the inventory is lower than the safety stock level, a replenishment strategy is executed and the procurement department is notified to purchase. When goods are slow to sell, the inventory turnover rate is further analyzed, and the inventory turnover cycle is calculated using the turnover rate calculation formula. When the sales cycle is longer than six months, discounts and promotions are implemented and cross-regional allocation is carried out through price optimization algorithms to reduce inventory pressure.

[0021] The specific formula for calculating turnover rate is as follows:

[0022]

[0023] Where S is the turnover period of a certain commodity; f is the current inventory of a certain commodity; w is the average monthly sales volume of a certain commodity; and p represents the turnover rate adjustment coefficient.

[0024] The logistics and delivery management module is responsible for improving logistics efficiency. It uses order volume and warehouse status data, and adjusts delivery routes based on real-time traffic and weather changes. It tracks the location, speed, and estimated arrival time of delivery vehicles in real time using onboard positioning devices, updates delivery status in real time, and provides customers and relevant personnel with delivery time predictions and vehicle location information. The logistics and delivery management module analyzes historical delivery data, uses regression analysis to identify the optimal delivery cycle, optimizes delivery plans, and flexibly adjusts the delivery cycle according to product characteristics, destination, and delivery time period.

[0025] The visualization support module displays key performance indicator data, including supplier on-time delivery rate, inventory turnover rate, and logistics delivery time, through dynamic charts and bar graphs, thereby helping managers understand the operational status.

[0026] This invention provides a supply chain management platform based on big data. Compared with existing technologies, it has the following advantages:

[0027] This invention uses a predictive analysis module to perform multi-dimensional demand analysis, combining historical sales data and market trends to predict demand fluctuations for different product categories, regions, and sales channels. Based on this prediction, the inventory management module can optimize inventory levels, avoid stockouts or excess inventory, and reduce inventory pressure and improve inventory turnover through replenishment strategies and promotional methods for slow-moving products.

[0028] Through the collaboration of a data acquisition module, a supplier management module, and a logistics and distribution management module, this invention enables the platform to track data at each stage of the supply chain, including production progress, inventory status, and transportation status, thereby improving the responsiveness and transparency of the entire supply chain. Real-time updates of supplier ratings and delivery schedules help to manage supplier relationships more effectively, ensure timely delivery, improve logistics efficiency, increase customer satisfaction, and reduce logistics costs. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the principle framework of the present invention. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] Please see Figure 1 This application provides a big data-based supply chain management platform, including a data acquisition module, a data cleaning and preprocessing module, a predictive analysis module, a supplier management module, an inventory management module, a logistics and distribution management module, and a visualization support module;

[0032] The data acquisition module collects key data from various platforms. It uses API interfaces and data crawlers to collect data in batches on a regular basis and transmits it to the data cleaning and preprocessing module.

[0033] The data cleaning and preprocessing module includes a data cleaning unit and a preprocessing unit. The data cleaning unit identifies and deletes duplicate records using the SHA-256 hash algorithm. The preprocessing unit fills in missing parts of numerical data with the mean and processes categorical data through label encoding. The processed data is then transmitted to the predictive analysis module.

[0034] The predictive analytics module uses the processed data to build predictive models and uses LSTM models to capture trends, seasonality, and periodicity in long-term time series. This module conducts multi-dimensional demand analysis of categories, regions, and channels through historical sales data, market trends, and customer feedback. It also forecasts demand by analyzing seasonal changes through time series analysis, uses geographic weighted regression for regional demand forecasting, and uses cross-validation to forecast channel-level demand.

[0035] The supplier management module provides a supplier communication window, allowing suppliers to synchronize production progress and delivery information, and provides a shared file and document management window. It also adjusts supplier ratings through a weighted algorithm and displays the ratings in the supplier communication window.

[0036] The inventory management module calculates the demand for the next cycle based on the forecasting model of the forecasting analysis module and generates a procurement plan. Based on the forecasted demand data, the inventory management module adjusts the inventory level and executes a replenishment strategy when the inventory is below the safety stock. When goods are not selling well, the module analyzes the turnover cycle of the goods and uses price optimization algorithms to carry out discount promotions and cross-regional allocation to reduce inventory pressure.

[0037] The logistics and delivery management module tracks the location, speed, and estimated arrival time of delivery vehicles in real time through on-board positioning devices, and provides delivery time forecasts and vehicle location information. It also optimizes delivery cycles and delivery plans through historical delivery data and regression analysis.

[0038] The visualization support module displays key performance indicator data through dynamic charts, helping managers understand the operational status.

[0039] The data acquisition module is responsible for collecting data from various sources, including: sales volume, sales time, and sales channels from POS, CRM, and e-commerce platforms; production plans, actual data, equipment status, and raw material consumption from MES and ERP systems; transportation status, transportation time, and transportation costs from TMS logistics systems; delivery status and contract fulfillment from suppliers on supply chain management platforms; and market data from social media and industry reports, including price fluctuations and consumer trends. The module uses API interfaces and web crawlers to periodically collect data from multiple platforms in batches, transmitting the collected data to the data cleaning and preprocessing module.

[0040] The data cleaning and preprocessing module improves the quality and usability of the collected data. It includes a data cleaning unit and a preprocessing unit. The data cleaning unit identifies and deletes duplicate records using the SHA-256 hash algorithm. The preprocessing unit fills in missing parts of numerical data with the mean, then uses the Z-Score method to identify and delete outliers in the data. The commonly used threshold is set to ±3 standard deviations. Finally, Min-Max standardization is used to map the data to the [0,1] interval, and LabelEncoding is used to process categorical data, unifying the data format and standardizing the data. The processed data of each category is then transmitted to the predictive analysis module.

[0041] The SHA-256 hash algorithm is used to identify and remove duplicate records. This process determines whether a record is a duplicate by converting data into a fixed-length hash value. Python's hashlib library provides an implementation of the SHA-256 algorithm to generate a hash value for each record, ensuring the uniqueness of each record.

[0042] import hashlib

[0043] def hash_record(record):

[0044] return hashl ib.sha256(record.encode('utf-8')).hexdigest()

[0045] The data preprocessing unit uses the mean imputation method to fill in missing values ​​in numerical data to make the data as complete as possible; Python's pandas library provides the `filllnna()` method, which conveniently performs mean imputation on missing values.

[0046] import pandas as pd

[0047] data=pd.DataFrame({'column1':[1,2,None,4,5]})

[0048] data['column1'].fil lna(data['column1'].mean(),inplace=True)

[0049] The data preprocessing unit maps data to the range [0,1], which is suitable for situations in machine learning models where a uniform data scale is required; Python's scikit-learn library provides MinMaxScaler to directly standardize data.

[0050] from sklearn.preprocessing import MinMaxScaler

[0051] import numpy as np

[0052] data=np.array([[1],[2],[3],[4],[5]])

[0053] scaler = MinMaxScaler()

[0054] normalized_data=scaler.fit_transform(data)

[0055] print(normalized_data)

[0056] The predictive analysis module uses the time series analysis methods ARIMA and ExponentialSmoothing to process the data. ARIMA is used to capture short-term trends and seasonal fluctuations, while ExponentialSmoothing is used to smooth the data and predict future demand fluctuations. The LSTM model is used to model the trends, seasonality and periodicity in long-term time series, which is suitable for demand data with complex time dependencies.

[0057] Exponential smoothing is a method for smoothing time series data, effectively capturing trends and seasonal variations. This method is based on a weighted average, assigning greater weight to more recent data points. Python's statsmodels library provides the ExponentialSmoothing class for smoothing data.

[0058] from statsmodels.tsa.holtwinters import ExponentialSmoothing

[0059] #data represents time series data

[0060] model=ExponentialSmoothing(data, trend='add', seasonal='add', seasonal_periods=12)

[0061] model_fit = model.fit()

[0062] forecast = model_fit.forecast(steps = 10) # Predict the next 10 steps

[0063] print(forecast)

[0064] LSTM can remember long-term information and effectively process long-term series data; Python's Keras or TensorFlow libraries provide implementations of LSTM models for training and predicting time series data.

[0065] from tensorflow.keras.models import Sequential

[0066] from tensorflow.keras.layers import LSTM,Dense

[0067] import numpy as np

[0068] #data represents the processed time series data.

[0069] data=np.array(data).reshape((len(data),1))

[0070] # Define the LSTM model

[0071] model = Sequential()

[0072] model.add(LSTM(50,return_sequences=True,input_shape=(data.shape[1],1)))

[0073] model.add(LSTM(50,return_sequences=False))

[0074] model.add(Dense(1)) # Output layer

[0075] model.compile(optimizer='adam',loss='mean_squared_error')

[0076] #Training the model

[0077] model.fit(data,data,epochs=20,batch_size=32)

[0078] #predict

[0079] forecast=model.predict(data)

[0080] print(forecast)

[0081] The predictive analysis module performs multi-dimensional demand analysis based on historical sales data, market trends, and customer feedback, considering product category, region, and channel. It also employs time series analysis to decompose the seasonal variations of each product category. For regional demand forecasting, it uses Geographically Weighted Regression (GWR) to analyze demand characteristics in different regions and combines these with regional consumption characteristics and economic conditions to create customized forecasts. Furthermore, the predictive analysis module analyzes demand fluctuations across different sales channels and uses cross-validation to predict channel-level demand.

[0082] In the data processing, the ARIMA model is first used to differencing the historical data to eliminate trend and seasonal components. Then, Exponential Smoothing is used to smooth the stationary data in order to capture future demand fluctuation trends. The LSTM model is trained on historical demand data to capture long-term time dependencies, thereby predicting demand fluctuations and cycles and providing accurate demand forecast results.

[0083] Furthermore, this invention uses a predictive analysis module to perform multi-dimensional demand analysis, combining historical sales data and market trends to predict demand fluctuations for different product categories, regions, and sales channels. Based on this prediction, the inventory management module can optimize inventory levels, avoid stockouts or excess inventory, and reduce inventory pressure and improve inventory turnover through replenishment strategies and promotional methods for slow-moving products.

[0084] The supplier management module is responsible for managing suppliers and improving supply chain efficiency. It provides a supplier communication window to synchronize production progress and delivery information, and a shared file and document management window to manage contracts, orders, quality reports, and documents. It tracks key performance indicators of suppliers, such as delivery time, product quality, and customer service level. Based on historical data, it adjusts supplier scores through a weighted algorithm and displays the supplier scores in the supplier communication window.

[0085] For key performance indicators (KPIs) of suppliers, such as delivery time, product quality, and customer service level, the overall score of suppliers is adjusted by setting different weights. The numpy and pandas libraries in Python are used for weight allocation and calculation. The implementation of the weighted model usually requires the combination of data tables and weighting coefficients for calculation. Assuming that each supplier's score is based on multiple dimensions (delivery time, product quality, customer service, etc.), the overall score is calculated using a weighted algorithm. The weights are configured by the system and can be adjusted according to business needs.

[0086]

[0087]

[0088] Where df is a DataFrame of supplier ratings, which includes multiple dimensions (delivery time, product quality, customer service, etc.);

[0089] weights is a dictionary representing the weights of each rating dimension;

[0090] By multiplying the score of each dimension by its corresponding weight, a weighted score for each supplier is obtained.

[0091] The inventory management module calculates the demand for the next cycle based on the predictive model in the predictive analysis module and generates a procurement plan. It adjusts the inventory according to the predicted demand data. When the inventory is lower than the safety stock level, a replenishment strategy is executed and the procurement department is notified to purchase. When goods are slow to sell, the inventory turnover rate is further analyzed, and the inventory turnover cycle is calculated using the turnover rate calculation formula. When the sales cycle is longer than six months, discounts and promotions are implemented and cross-regional allocation is carried out through price optimization algorithms to reduce inventory pressure.

[0092] The specific formula for calculating turnover rate is as follows:

[0093]

[0094] Where S is the turnover period of a certain commodity; f is the current inventory of a certain commodity; w is the average monthly sales volume of a certain commodity; and p represents the turnover rate adjustment coefficient.

[0095] For slow-moving goods, when the turnover cycle exceeds six months, the platform will use a price optimization algorithm to offer discounts and promotions. Price optimization typically relies on a demand elasticity model, adjusting prices to maximize sales volume. Python's scikit-learn library is used to implement regression analysis, calculating the relationship between demand and price. A linear regression model is used to fit the relationship between product price and sales volume, thereby calculating the optimal pricing.

[0096] from sklearn.linear_model import LinearRegression

[0097] model = LinearRegression()

[0098] model.fit(price_data, sales_data) # price_data is the price, sales_data is the sales volume

[0099] predicted_sales = model.predict(new_price_data) # Predict sales volume at different prices

[0100] The logistics and delivery management module is responsible for improving logistics efficiency. It uses order volume and warehouse status data, and adjusts delivery routes based on real-time traffic and weather changes. It tracks the location, speed, and estimated arrival time of delivery vehicles in real time using onboard positioning devices, updates delivery status in real time, and provides customers and relevant personnel with delivery time predictions and vehicle location information. The logistics and delivery management module analyzes historical delivery data, uses regression analysis to identify the optimal delivery cycle, optimizes delivery plans, and flexibly adjusts the delivery cycle according to product characteristics, destination, and delivery time period.

[0101] The visualization support module displays key performance indicator data, including supplier on-time delivery rate, inventory turnover rate, and logistics delivery time, through dynamic charts and bar graphs, thereby helping managers understand the operational status.

[0102] Furthermore, through the collaboration of the data acquisition module, supplier management module, and logistics and distribution management module, the platform can track data at each stage of the supply chain, including production progress, inventory status, and transportation status, thereby improving the responsiveness and transparency of the entire supply chain. Real-time updates of supplier ratings and delivery progress help to manage supplier relationships more effectively, ensure timely delivery, improve logistics efficiency, increase customer satisfaction, and reduce logistics costs.

[0103] Specific workflow:

[0104] S1: The data acquisition module collects data from various platforms in batches on a regular basis through API interfaces and data crawlers, and transmits the collected data to the data cleaning and preprocessing module. The data includes sales data from POS, CRM, and e-commerce platforms, factory data from MES and ERP, and logistics data from TMS.

[0105] S2: The data cleaning and preprocessing module uses the hash algorithm SHA-256 to identify and delete duplicate records through the data cleaning unit. The preprocessing unit fills the mean of numerical data, uses the Z-Score method to identify and delete outliers, and then uses Min-Max standardization and LabelEncoding to process categorical data to standardize the data before transmitting it to the predictive analysis module.

[0106] S3: The predictive analytics module uses the time series analysis methods ARIMA and ExponentialSmoothing to process short-term demand data, uses the LSTM model to capture long-term demand trends, and uses geographic weighted regression analysis to predict regional demand. It also uses cross-validation to predict channel-level demand and conducts multi-dimensional demand analysis based on historical data, market trends, and customer feedback.

[0107] S4: The supplier management module provides a supplier communication window to display production progress and delivery information, and provides a shared file management window to support the management of contracts, orders and quality reports. It also adjusts supplier ratings through a weighted algorithm and displays them in the communication window.

[0108] S5: The inventory management module calculates the demand for the next cycle based on the predictive model of the predictive analysis module and generates a procurement plan. The inventory management module manages the inventory level. When the inventory is lower than the safety stock, it executes a replenishment strategy to notify the procurement department to make purchases. When goods are slow to sell, it uses price optimization algorithms to promote sales and allocate goods across regions to reduce inventory pressure.

[0109] S6: The logistics and delivery management module tracks the location, speed, and estimated arrival time of delivery vehicles in real time through vehicle positioning devices, adjusts delivery routes in conjunction with real-time traffic and weather information, and provides customers with delivery time forecasts and vehicle location information. It also optimizes delivery cycles and delivery plans through regression analysis.

[0110] S7: The visualization support module displays key performance indicators such as supplier on-time delivery rate, inventory turnover rate, and logistics delivery time through dynamic charts and bar graphs, helping managers understand the operational status.

[0111] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.

[0112] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A supply chain management platform based on big data, characterized in that, include: The system includes a data acquisition module, a data cleaning and preprocessing module, a predictive analysis module, a supplier management module, an inventory management module, a logistics and distribution management module, and a visualization support module. The data acquisition module collects key data from various platforms. The data acquisition module collects data in batches on a regular basis through API interfaces and data crawlers and transmits it to the data cleaning and preprocessing module. The data cleaning and preprocessing module includes a data cleaning unit and a preprocessing unit. The data cleaning unit identifies and deletes duplicate records using the SHA-256 hash algorithm. The preprocessing unit fills in missing parts of numerical data with the mean and processes categorical data through label encoding. The processed data is then transmitted to the predictive analysis module. The predictive analytics module uses the processed data to build predictive models and uses LSTM models to capture trends, seasonality, and periodicity in long-term time series. This module conducts multi-dimensional demand analysis of categories, regions, and channels through historical sales data, market trends, and customer feedback. It also forecasts demand by analyzing seasonal changes through time series analysis, uses geographic weighted regression for regional demand forecasting, and uses cross-validation to forecast channel-level demand. The supplier management module provides a supplier communication window, allowing suppliers to synchronize production progress and delivery information, and provides a shared file and document management window. It also adjusts supplier ratings through a weighted algorithm and displays the ratings in the supplier communication window. The inventory management module calculates the demand for the next cycle based on the forecasting model of the forecasting analysis module and generates a procurement plan. Based on the forecasted demand data, the inventory management module adjusts the inventory level and executes a replenishment strategy when the inventory is below the safety stock. When goods are not selling well, the module analyzes the turnover cycle of the goods and uses price optimization algorithms to carry out discount promotions and cross-regional allocation to reduce inventory pressure. The logistics and delivery management module tracks the location, speed, and estimated arrival time of delivery vehicles in real time through on-board positioning devices, and provides delivery time forecasts and vehicle location information. It also optimizes delivery cycles and delivery plans through historical delivery data and regression analysis. The visualization support module displays key performance indicator data through dynamic charts, helping managers understand the operational status.

2. The supply chain management platform based on big data according to claim 1, characterized in that, The data acquisition module is responsible for collecting data, including sales volume, sales time, and sales channels from POS, CRM, and e-commerce platforms; production plans, actual data, equipment status, and raw material consumption from MES and ERP factory systems; transportation status, transportation time, and transportation costs from TMS logistics systems; delivery status and contract fulfillment from supply chain management platforms; and market data from social media and industry reports, including price fluctuations and consumer trends. The data acquisition module collects data from multiple platforms in batches on a regular basis through API interfaces and data crawlers, and then transmits the collected data to the data cleaning and preprocessing module.

3. The supply chain management platform based on big data according to claim 1, characterized in that, The data cleaning and preprocessing module improves the quality and usability of the collected data. It includes a data cleaning unit and a preprocessing unit. The data cleaning unit identifies and deletes duplicate records using the SHA-256 hash algorithm. The preprocessing unit fills in missing parts of numerical data with the mean, then uses the Z-Score method to identify and delete outliers in the data. The commonly used threshold is set to ±3 standard deviations. Finally, Min-Max standardization is used to map the data to the [0,1] interval, and LabelEncoding is used to process categorical data, unifying the data format and standardizing the data. The processed data of each category is then transmitted to the predictive analysis module.

4. The supply chain management platform based on big data according to claim 1, characterized in that, The predictive analysis module uses the time series analysis methods ARIMA and ExponentialSmoothing to process the data. ARIMA is used to capture short-term trends and seasonal fluctuations, while ExponentialSmoothing is used to smooth the data and predict future demand fluctuations. The LSTM model is used to model the trends, seasonality and periodicity in long-term time series, which is suitable for demand data with complex time dependencies. The predictive analysis module performs multi-dimensional demand analysis based on historical sales data, market trends, and customer feedback, considering product category, region, and channel. It also employs time series analysis to decompose the seasonal variations of each product category. For regional demand forecasting, it uses Geographically Weighted Regression (GWR) to analyze demand characteristics in different regions and combines these with regional consumption characteristics and economic conditions to create customized forecasts. Furthermore, the predictive analysis module analyzes demand fluctuations across different sales channels and uses cross-validation to predict channel-level demand. In the data processing, the ARIMA model is first used to differencing the historical data to eliminate trend and seasonal components. Then, Exponential Smoothing is used to smooth the stationary data in order to capture future demand fluctuation trends. The LSTM model is trained on historical demand data to capture long-term time dependencies, thereby predicting demand fluctuations and cycles and providing accurate demand forecast results.

5. A supply chain management platform based on big data according to claim 1, characterized in that, The supplier management module is responsible for managing suppliers and improving supply chain efficiency. It provides a supplier communication window to synchronize production progress and delivery information, and a shared file and document management window to manage contracts, orders, quality reports, and documents. It tracks key performance indicators of suppliers, such as delivery time, product quality, and customer service level. Based on historical data, it adjusts supplier scores through a weighted algorithm and displays the supplier scores in the supplier communication window.

6. A supply chain management platform based on big data according to claim 1, characterized in that, The inventory management module calculates the demand for the next cycle based on the predictive model in the predictive analysis module and generates a procurement plan. It adjusts the inventory according to the predicted demand data. When the inventory is lower than the safety stock level, a replenishment strategy is executed and the procurement department is notified to purchase. When goods are slow to sell, the inventory turnover rate is further analyzed, and the inventory turnover cycle is calculated using the turnover rate calculation formula. When the sales cycle is longer than six months, discounts and promotions are implemented and cross-regional allocation is carried out through price optimization algorithms to reduce inventory pressure. The specific formula for calculating turnover rate is as follows: Where S is the turnover period of a certain commodity; f is the current inventory of a certain commodity; w is the average monthly sales volume of a certain commodity; and p represents the turnover rate adjustment coefficient.

7. A supply chain management platform based on big data according to claim 1, characterized in that, The logistics and delivery management module is responsible for improving logistics efficiency. It uses order volume and warehouse status data, and adjusts delivery routes based on real-time traffic and weather changes. It tracks the location, speed, and estimated arrival time of delivery vehicles in real time using onboard positioning devices, updates delivery status in real time, and provides customers and relevant personnel with delivery time predictions and vehicle location information. The logistics and delivery management module analyzes historical delivery data, uses regression analysis to identify the optimal delivery cycle, optimizes delivery plans, and flexibly adjusts the delivery cycle according to product characteristics, destination, and delivery time period.

8. A supply chain management platform based on big data according to claim 1, characterized in that, The visualization support module displays key performance indicator data, including supplier on-time delivery rate, inventory turnover rate, and logistics delivery time, through dynamic charts and bar graphs, thereby helping managers understand the operational status.

Images