Rapid extraction method and device for enterprise sales data and medium
By sorting, preprocessing and combining enterprise sales data, combining text query and chart query, the problem of inefficiency of enterprise decision makers when looking for sales data in the database is solved, and faster and more accurate data acquisition is achieved.
Patent Information
- Application Number
- CN202411905920.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-23
- Publication Date
- 2025-04-11
AI Technical Summary
When corporate decision makers obtain corporate retail sales data in the database, due to the huge and disordered data, the search speed is slow and the efficiency is low.
By classifying and pre-processing the enterprise sales data obtained from different data sources, forming a joint data set, and providing text query and chart query methods, determining the data analysis algorithm based on user query requests, processing and formal conversion of the data.
Improve data quality and consistency, enhance the correlation between data, and enable users to obtain the required information more quickly and accurately to meet the needs and preferences of different users.
Smart Images
Figure CN119807282A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of data processing technology, and in particular to a method, device and medium for quickly extracting enterprise sales data. Background Art
[0002] In today's rapidly changing business era, corporate sales data has become an important basis for companies to formulate sales strategies and evaluate market performance.
[0003] Enterprise sales data often comes from multiple heterogeneous systems, such as sales management systems, customer relationship management systems, and e-commerce platforms. Especially for some large retail enterprises, which have many retail stores, data needs to be updated daily, and a large amount of sales data is generated every month and every year. The amount of this data is huge, and there is a lack of unified data architecture and classification standards, which makes data maintenance and data query extremely cumbersome.
[0004] In the prior art, enterprise decision makers often obtain enterprise retail sales data from a database as needed. However, due to the huge amount and disorder of the data, the search speed is slow and the efficiency is low when extracting the required data. Summary of the invention
[0005] The embodiments of the present application provide a method, device and medium for optimizing enterprise decision making, which are used to solve the following technical problems: enterprise decision makers obtain enterprise retail sales data from a database as needed. Due to the huge amount and disorder of the data, the search speed is slow and the efficiency is low when extracting the required data.
[0006] The present application embodiment adopts the following technical solutions:
[0007] The embodiment of the present application provides a method for optimizing enterprise decision making, including: classifying and storing enterprise sales data obtained from different data sources; preprocessing the classified and stored enterprise sales data, and associating and merging the preprocessed enterprise sales data to obtain an associated and merged data set; after receiving a query request sent by a user, extracting data from the associated and merged data set by using a query method corresponding to the query request; wherein the query method includes text query and chart query; determining a data analysis algorithm based on the query request, and processing the extracted enterprise sales data based on the data analysis algorithm and enterprise sales data keywords; converting the processed enterprise sales data into a form and sending it to the user end.
[0008] The embodiments of the present application can clean and organize data through preprocessing, remove redundant and erroneous information, and improve data quality. Through association merging, data from different data sources can be integrated into a unified data set, so that the relevance and consistency between data are strengthened. Users can more easily obtain comprehensive and accurate information when querying and analyzing data. Text query and chart query are provided to meet the needs and preferences of different users. Text query is suitable for users who need to accurately find specific information, while chart query is more suitable for users who need to intuitively understand data trends and relationships, improving the efficiency and satisfaction of queries. Determine the data analysis algorithm based on the query request to ensure that the processed data better meets the needs of users.
[0009] In one implementation of the present application, data is extracted from an associated merged data set through a query method corresponding to a query request, specifically including: when the query method corresponding to the query request is a text query, determining text keywords in the query request, extracting data from the associated merged data set based on the text keywords, and summarizing the acquired data in a tabular form; when the query method corresponding to the query request is a chart query, acquiring chart structure information corresponding to the chart; wherein the chart structure information includes at least one of a chart type, a chart title, and chart axis information; extracting data from the associated merged data set based on the chart structure information, and filling the chart with content based on the extracted data.
[0010] In one implementation of the present application, the classified and stored enterprise sales data are preprocessed, and the preprocessed enterprise sales data are associated and merged to obtain an associated and merged data set, which specifically includes: data cleaning of the classified and stored enterprise sales data; format classification of the cleaned enterprise sales data, and format unification of enterprise sales data of different format categories; based on preset association rules and data sources, the enterprise sales data with unified formats are associated and merged to obtain an associated and merged data set.
[0011] In one implementation of the present application, a data analysis algorithm is determined based on a query request, and the extracted enterprise sales data is processed based on the data analysis algorithm and the enterprise sales data keywords, specifically including: determining the enterprise sales data keywords based on a preset keyword table; determining the data arrangement type based on the enterprise sales data keywords, and grouping and arranging the enterprise sales data according to the data arrangement type; determining the data analysis type based on the query request; wherein the data analysis type includes at least one of data summation, data comparison, and data grouping; determining the data analysis algorithm based on the data analysis type, and processing the grouped and arranged enterprise sales data based on the data analysis algorithm.
[0012] In one implementation of the present application, the processed enterprise sales data is converted into a format, specifically including: converting the received enterprise sales data into a natural language format or a visual report format; wherein the received enterprise sales data is structured data.
[0013] In one implementation of the present application, after receiving a query request sent by a user, the method also includes: matching in a historical database based on the query request to obtain historical query text information; wherein the historical query text information includes at least one of historical question records, the department to which the user belongs, and the current business scenario; based on the historical query text information, the query request, and a first preset deep learning model, expanding the query request information to obtain multidimensional query data; and, based on the historical query text information, the query request, and a second preset deep learning model, recommending data analysis tasks and self-service analysis operation tasks for the query request.
[0014] In one implementation of the present application, after matching in a historical database based on a query request to obtain historical query text information, the method also includes: determining a corresponding machine learning model when the query request meets a preset analysis difficulty condition; training the machine learning model based on the historical query text information corresponding to the query request to obtain the required machine model; and analyzing the query request based on the required machine model.
[0015] In one implementation of the present application, after the enterprise sales data obtained from different data sources are classified and stored, the method also includes: identifying sensitive data of the stored enterprise sales data, encrypting the identified sensitive data, and transmitting the enterprise sales data to the data processing layer via the SSL / TLS encryption protocol; dividing the enterprise sales data based on the data type; wherein the data type includes at least one of a role data type, a department data type, and a business requirement data type; and assigning different data access rights to users based on the data division results.
[0016] An embodiment of the present application provides an enterprise decision optimization device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can: classify and store enterprise sales data acquired from different data sources; preprocess the classified and stored enterprise sales data, and associate and merge the preprocessed enterprise sales data to obtain an associated and merged data set; after receiving a query request sent by a user, extract data from the associated and merged data set through a query method corresponding to the query request; wherein the query method includes text query and chart query; determine a data analysis algorithm based on the query request, and process the extracted enterprise sales data based on the data analysis algorithm and enterprise sales data keywords; convert the processed enterprise sales data into a format and send it to a user end.
[0017] A non-volatile computer storage medium provided in an embodiment of the present application stores computer executable instructions, wherein the computer executable instructions are configured to: classify and store enterprise sales data obtained from different data sources; preprocess the classified and stored enterprise sales data, and associate and merge the preprocessed enterprise sales data to obtain an associated and merged data set; after receiving a query request sent by a user, extract data from the associated and merged data set through a query method corresponding to the query request; wherein the query method includes a text query and a chart query; determine a data analysis algorithm based on the query request, and process the extracted enterprise sales data based on the data analysis algorithm and enterprise sales data keywords; convert the processed enterprise sales data into a format and send it to a user end.
[0018] At least one of the above technical solutions adopted in the embodiments of the present application can achieve the following beneficial effects: the embodiments of the present application can clean and organize data through preprocessing steps, remove redundant and erroneous information, and improve data quality. Through association merging, data from different data sources can be integrated into a unified data set, so that the relevance and consistency between data are strengthened. Users can more easily obtain comprehensive and accurate information when querying and analyzing data. Text query and chart query are provided to meet the needs and preferences of different users. Text query is suitable for users who need to find specific information accurately, while chart query is more suitable for users who need to intuitively understand data trends and relationships, improving query efficiency and satisfaction. Determine the data analysis algorithm based on the query request to ensure that the processed data better meets the needs of the user. BRIEF DESCRIPTION OF THE DRAWINGS
[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. In the drawings:
[0020] Figure 1 A flowchart for quickly extracting enterprise sales data provided by an embodiment of the present application;
[0021] Figure 2 A flow chart of enterprise sales data processing provided in an embodiment of the present application;
[0022] Figure 3 A ChatBI interaction example diagram provided in an embodiment of the present application;
[0023] Figure 4 A schematic diagram of the structure of a device for quickly extracting enterprise sales data provided in an embodiment of the present application. DETAILED DESCRIPTION
[0024] The embodiments of the present application provide a method, device and medium for quickly extracting enterprise sales data.
[0025] In order to enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of them. Based on the embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this application.
[0026] The technical solution proposed in the embodiments of the present application is described in detail below with reference to the accompanying drawings.
[0027] In one embodiment of the present application, a method for rapidly extracting enterprise sales data is performed by a system for rapidly extracting enterprise sales data, and the system includes a data storage layer, a data processing layer, and a ChatBI interaction layer.
[0028] The data storage layer is used to store various structured and unstructured data within the enterprise, including but not limited to business transaction data, customer relationship management data, market research data, log data, etc. Data storage can adopt a variety of storage technologies such as relational databases (such as MySQL, Oracle, etc.), non-relational databases (such as MongoDB, Elasticsearch, etc.) and data lakes (such as data lakes built with Hadoop distributed file systems). Reasonable selection and configuration should be made according to the characteristics of the data and usage scenarios to ensure efficient storage, retrieval and management of data.
[0029] The data processing layer is responsible for pre-processing operations such as cleaning, conversion, and integration of the data in the storage layer, and converting the raw data into a format suitable for analysis. At the same time, this layer also contains a data analysis engine that can execute various data analysis algorithms and tasks, such as data mining, statistical analysis, machine learning model training and prediction, etc. The data analysis engine can be built based on an open source framework (such as Spark, TensorFlow, etc.), or it can use commercial data analysis software products with strong computing power and scalability to meet the analysis needs of large-scale data.
[0030] The ChatBI interaction layer is the main interface for users to interact with the system. It is implemented based on natural language processing technology. Users can ask the system various questions related to enterprise data through text input boxes or voice input, such as "How is the sales performance this month?" "Which customers have a higher risk of churn?" etc. After receiving the user input, the ChatBI interaction layer first parses and understands the natural language, extracts key information and intent, and then converts it into query statements or task instructions that can be recognized by the data analysis engine, and passes it to the data processing layer for processing.
[0031] The self-service analysis layer provides self-service analysis tools and environments for users who have certain data analysis capabilities or want to explore data in depth. This layer provides a visual data exploration interface, where users can filter, group, aggregate, draw charts, and perform other operations on data by dragging, clicking, and other operations, without having to write complex code to perform data analysis. At the same time, the self-service analysis layer is interconnected with the ChatBI interaction layer. When users encounter problems during the self-service analysis process or need further in-depth analysis, they can switch to the ChatBI interaction mode at any time to ask questions to the system and get help and guidance.
[0032] Figure 1 A flowchart for quickly extracting enterprise sales data provided by an embodiment of the present application is as follows: Figure 1 As shown, the method for quickly extracting enterprise sales data includes the following steps:
[0033] S101. Classify and store the enterprise sales data obtained from different data sources.
[0034] In one embodiment of the present application, the system collects data from various business systems within the enterprise (such as ERP system, CRM system, e-commerce platform, etc.) and external data sources (such as industry reports, social media data, etc.) regularly or in real time through multiple data source access methods, and stores it in the data storage layer. During the data storage process, the data is classified and stored according to its type, source and purpose, and a reasonable data directory and index structure are established to facilitate subsequent data management and query.
[0035] S102: Preprocess the classified and stored enterprise sales data, and associate and merge the preprocessed enterprise sales data to obtain an associated and merged data set.
[0036] In one embodiment of the present application, data cleaning is performed on the classified and stored enterprise sales data. The cleaned enterprise sales data is format-classified, and the enterprise sales data of different format categories are format-unified. Based on the preset association rules, the enterprise sales data of the unified format are associated and merged to obtain an associated and merged data set.
[0037] Specifically, after the data processing layer in the embodiment of the present application reads data from the data storage layer, it first performs a data cleaning operation to remove noise, duplicate data, erroneous data, etc. in the data. For example, abnormal transaction amounts in business transaction data (such as order amounts that are obviously beyond the normal range) are identified and corrected or marked. Then a data conversion operation is performed to uniformly convert data of different formats and different encodings into a standard format, such as unifying the date format to "YYYY-MM-DD", converting string type values to digital types, etc. Finally, a data integration operation is performed to associate and merge data from multiple data sources according to certain rules, such as associating the basic information of customers in the CRM system with transaction records on the e-commerce platform to form complete customer portrait data.
[0038] S103: After receiving the query request sent by the user, extract data from the associated merged data set using a query method corresponding to the query request.
[0039] In one embodiment of the present application, when the query method corresponding to the query request is a text query, the text keywords in the query request are determined, data is extracted from the associated merged data set based on the text keywords, and the acquired data is summarized in a table. When the query method corresponding to the query request is a chart query, the chart structure information corresponding to the chart is acquired; wherein the chart structure information includes at least one of the chart type, the chart title, and the chart axis information. Data is extracted from the associated merged data set based on the chart structure information, and the chart is filled with content based on the extracted data. Among them, the query methods include text query and chart query.
[0040] Specifically, when the query method corresponding to the query request is a text query, a preset data analysis engine is used to extract keywords from the query request input by the user. Based on the extracted keywords, keyword matching is performed in the associated merged data set to extract the required data based on the matching results.
[0041] A chart query refers to a user request to display data in the form of a chart. Users usually specify the chart type, such as a bar chart, line chart, etc., the title, and axis information, such as the X-axis represents time and the Y-axis represents sales. The system needs to extract the corresponding data from the associated merged data set based on this information and generate the chart specified by the user. Specifically, when the query method corresponding to the query request is a chart query, the chart type, title, and axis information are extracted from the query request, the corresponding data is extracted from the data set based on the chart structure information, the extracted data is used to fill the chart, and the chart requested by the user is generated.
[0042] S104: Determine a data analysis algorithm based on the query request, and process the extracted enterprise sales data based on the data analysis algorithm and enterprise sales data keywords.
[0043] In one embodiment of the present application, enterprise sales data keywords are determined based on a preset keyword table. Data arrangement types are determined based on enterprise sales data keywords, so that enterprise sales data are grouped and arranged according to the data arrangement types. Data analysis types are determined based on query requests; wherein the data analysis types include at least one of data summation, data comparison, and data grouping. Data analysis algorithms are determined based on the data analysis types, so that the grouped and arranged enterprise sales data are processed based on the data analysis algorithms.
[0044] Specifically, when the ChatBI interaction layer receives a user's query request, it passes it to the data processing layer. The data analysis engine of the data processing layer extracts keywords based on the request content, matches the extracted keywords in the associated merged data set, extracts the required relevant data, and uses the corresponding analysis algorithm for analysis.
[0045] For example, if a user asks "What is the ranking of total sales in each region in the past year?", the data analysis engine will group and sum the sales data by region and sort them in descending order by total sales. The analysis results are returned to the ChatBI interaction layer in the form of structured data, which is then converted into natural language or visual reports and presented to the user.
[0046] The self-service analysis layer is associated with the ChatBI interaction layer. When a switching instruction is received, the self-service analysis layer is switched to the ChatBI interaction layer to analyze the question information input by the user based on the interaction mode of the ChatBI interaction layer.
[0047] Furthermore, the self-service analysis layer of the enterprise decision optimization system obtains the data operation instructions sent by the user, and according to the conditions set by the user, selects the data subset that meets the specific requirements from the huge data set, and groups the selected data according to one or more attributes to facilitate more in-depth analysis. Perform statistical calculations on the grouped data, such as sum, average, maximum value, minimum value, etc., to draw more detailed and accurate conclusions. The processed data is displayed in the form of charts, so that users can understand the data and analysis results more intuitively. The self-service analysis layer can directly obtain the required data from the data processing layer. The data processing layer is usually responsible for data cleaning, conversion and integration to ensure the accuracy and consistency of the data.
[0048] Furthermore, after acquiring the data, the self-service analysis layer will process and analyze the data accordingly according to the data operation instructions sent by the user. The self-service analysis layer is also associated with the ChatBI interaction layer.
[0049] The ChatBI interaction layer provides a more natural and intuitive way of interaction, allowing users to input questions or instructions through natural language or dialogue. When the self-service analysis layer receives a switching instruction, it switches control to the ChatBI interaction layer. In the ChatBI interaction layer, the system analyzes the question information entered by the user and returns corresponding results or suggestions. A switching instruction is a special instruction used to switch between the self-service analysis layer and the ChatBI interaction layer. This instruction is usually triggered by the user. For example, the user may want to enter a more complex query or instruction in a dialogue. At this time, the system will transfer control to the ChatBI interaction layer according to the switching instruction.
[0050] In one embodiment of the present application, after receiving a query request sent by a user, a match is performed in a historical database based on the query request to obtain historical query text information; wherein the historical query text information includes at least one of the historical question record, the department to which the user belongs, and the current business scenario. Based on the historical query text information, the query request, and the first preset deep learning model, the query request is expanded to obtain multidimensional query data. And, based on the historical query text information, the query request, and the second preset deep learning model, data analysis tasks and self-service analysis operation tasks are recommended for the query request.
[0051] Specifically, in the ChatBI interaction layer, the natural language processing algorithm is further optimized, and the deep learning model is used to improve the semantic understanding of user questions. In addition to understanding the user's direct question intention, it can also conduct a comprehensive analysis based on the user's historical question records, department, current business scenario and other contextual information to provide more accurate and personalized answers and data analysis suggestions.
[0052] For example, when a user in the marketing department asks about the "market performance of the product", the system can input the information such as the product series and market area that the user has previously paid attention to into the first preset deep learning model to output extended information. The training method of the first preset deep learning model is as follows: the historical user query statement, the product series and market area corresponding to the historical user query statement and other information are used as input samples, and the extended information corresponding to the input sample is used as an output sample to train the preset neural network model to obtain the first preset deep learning model. For example, the trained first preset deep learning model automatically provides detailed sales data, market share changes, customer feedback and other multi-dimensional analysis results about the specific product series in the relevant market area, rather than simply answering general market performance data about all products.
[0053] Furthermore, relevant data analysis tasks or self-service analysis operation steps can be automatically recommended based on the user's questions, guiding the user to further explore the data and improve the comprehensiveness and depth of decision-making. Specifically, the second preset deep learning model is used to perform data analysis and operation task recommendations on the query request input by the user. The process is: keyword extraction is performed on the query request input by the user, and the corresponding historical query text information is determined in the historical database based on the extracted keywords. The query request and the historical query text information are input into the second preset deep learning model to output the data analysis task recommendation and self-service analysis operation task recommendation corresponding to the query request.
[0054] Among them, the training method of the second preset deep learning model is: taking the historical user query request and the historical query text information corresponding to the historical user query request as input samples, and taking the historical data analysis task recommendation and the historical self-service analysis operation task recommendation corresponding to the input sample as output samples, inputting into the preset neural network model for model training, and obtaining the second preset deep learning model.
[0055] In one embodiment of the present application, when the query request meets the preset analysis difficulty condition, a corresponding machine learning model is determined. Based on the historical query text information corresponding to the query request, the machine learning model is trained to obtain the required machine model. The query request is analyzed based on the required machine model.
[0056] Specifically, the automated machine learning (AutoML) technology is introduced in the data analysis engine of the data processing layer. When the system receives complex analysis requirements from users or finds that some potential patterns in the data need to be further mined, the AutoML module can automatically select the appropriate machine learning algorithm (such as linear regression, decision tree, neural network, etc.), and automatically optimize and adjust the parameters of the algorithm to train the best model to meet the analysis requirements.
[0057] For example, when a user asks "predict the sales trend in the next three months", the AutoML module will automatically select the appropriate time series forecasting model (such as ARIMA model or time series forecasting model based on deep learning) according to the characteristics of historical sales data (such as data distribution, seasonal characteristics, trend, etc.), and determine the best model parameters through automated parameter adjustment and model evaluation process to generate accurate sales trend forecast results. This automated model training and optimization mechanism not only improves the efficiency and accuracy of data analysis, but also reduces the dependence on professional data scientists, enabling enterprises to respond more quickly to various complex business decision-making needs.
[0058] In one embodiment of the present application, sensitive data is identified for stored enterprise sales data, the identified sensitive data is encrypted, and the enterprise sales data is encrypted and transmitted via the SSL / TLS encryption protocol. The encrypted enterprise sales data is divided based on the data type; wherein the data type includes at least one of the role data type, department data type, and business requirement data type. Based on the data division results, different data access rights are assigned to users.
[0059] Specifically, data security and privacy protection measures are strengthened throughout the system architecture. In the data storage layer, data encryption technology is used to encrypt and store sensitive data, such as encrypting customers' personal identity information and financial data, to ensure the security of data on the storage medium. In the data transmission process, encryption protocols such as SSL / TLS are used to prevent data from being stolen or tampered with during network transmission. At the same time, a strict user rights management mechanism is implemented in the ChatBI interaction layer and self-service analysis layer, and different data access rights are assigned to users based on factors such as user roles, departments, and business needs.
[0060] For example, ordinary business personnel can only access data related to their own business, while financial personnel can access financial data, but only senior management can access sensitive data related to the company's core strategy under specific authorization. In addition, the system also has a data audit function that can record all users' operations on data so as to trace and investigate when a data security incident occurs.
[0061] S105: Convert the processed enterprise sales data into a different format and send it to the user end.
[0062] In one embodiment of the present application, the received enterprise sales data is converted into a natural language form or a visual report form; wherein the received enterprise sales data is structured data.
[0063] Figure 2 A flow chart of enterprise sales data processing provided in the embodiment of the present application, such as Figure 2 As shown in the figure, data is collected from various business systems within the enterprise and external data sources regularly or in real time, and stored in the data storage layer. After the data processing layer reads the data from the data storage layer, it first performs data cleaning operations to remove noise, duplicate data, erroneous data, etc. Finally, data integration operations are performed to associate and merge data from multiple data sources according to certain rules, generate corresponding reports for visual display, and send them to the user interface.
[0064] Figure 3 A ChatBI interaction example diagram provided in an embodiment of the present application, such as Figure 3As shown in the figure, when the ChatBI interaction layer receives the user's query request, it parses and understands the natural language, extracts key information and intent, and then converts it into a query statement or task instruction that the data analysis engine can recognize, and passes it to the data processing layer for processing. For example, if the user asks "the ranking of total sales in each region in the past year", the data analysis engine groups and sums the sales data according to the regional dimension, and sorts them in descending order according to the total sales. The analysis results are returned to the ChatBI interaction layer in the form of structured data, and the ChatBI interaction layer converts it into natural language or a visual report form to present it to the user.
[0065] Figure 4 This is a schematic diagram of a device for quickly extracting enterprise sales data provided by an embodiment of the present application. Figure 4 As shown, the enterprise decision optimization device includes: at least one processor; and, a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can: classify and store the enterprise sales data obtained from different data sources; preprocess the classified and stored enterprise sales data, and associate and merge the preprocessed enterprise sales data to obtain an associated and merged data set; after receiving a query request sent by a user, extract data from the associated and merged data set through a query method corresponding to the query request; wherein the query method includes text query and chart query; determine a data analysis algorithm based on the query request, and process the extracted enterprise sales data based on the data analysis algorithm and enterprise sales data keywords; convert the processed enterprise sales data into a form and send it to the user end.
[0066] A non-volatile computer storage medium provided in an embodiment of the present application stores computer executable instructions, wherein the computer executable instructions are configured to: classify and store enterprise sales data obtained from different data sources; preprocess the classified and stored enterprise sales data, and associate and merge the preprocessed enterprise sales data to obtain an associated and merged data set; after receiving a query request sent by a user, extract data from the associated and merged data set through a query method corresponding to the query request; wherein the query method includes a text query and a chart query; determine a data analysis algorithm based on the query request, and process the extracted enterprise sales data based on the data analysis algorithm and enterprise sales data keywords; convert the processed enterprise sales data into a format and send it to a user end.
[0067] Each embodiment in this application is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device, equipment, and non-volatile computer storage medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.
[0068] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the embodiments of the present application may have various changes and variations. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application.
Claims
1. A method for quickly extracting enterprise sales data, characterized in that: The method comprises: Categorize and store enterprise sales data obtained from different data sources; Preprocessing the classified and stored enterprise sales data, and associating and merging the preprocessed enterprise sales data to obtain an associated and merged data set; After receiving a query request sent by a user, extracting data from the associated merged data set by using a query method corresponding to the query request; wherein the query method includes text query and chart query; Determining a data analysis algorithm based on the query request, and processing the extracted enterprise sales data based on the data analysis algorithm and enterprise sales data keywords; The processed enterprise sales data is converted into a new format and sent to the user end.
2. The method for quickly extracting enterprise sales data according to claim 1, characterized in that: The extracting data from the associated merged data set by using the query method corresponding to the query request specifically includes: In the case where the query mode corresponding to the query request is a text query, determining text keywords in the query request, extracting data from the associated merged data set based on the text keywords, and summarizing the acquired data in a table form; In the case where the query mode corresponding to the query request is a chart query, obtaining chart structure information corresponding to the chart; wherein the chart structure information includes at least one of a chart type, a chart title, and chart axis information; Data is extracted from the associated merged data set based on the chart structure information, and content is filled into the chart based on the extracted data.
3. The method for quickly extracting enterprise sales data according to claim 1, characterized in that: The preprocessing of the classified and stored enterprise sales data and associating and merging the preprocessed enterprise sales data to obtain an associated and merged data set specifically includes: Performing data cleaning on the classified and stored enterprise sales data; Classifying the cleaned enterprise sales data in different formats and unifying the formats of the enterprise sales data in different formats; Based on the preset association rules and the data source, the enterprise sales data in a unified format are associated and merged to obtain the associated merged data set.
4. The method for quickly extracting enterprise sales data according to claim 1, characterized in that: The step of determining a data analysis algorithm based on the query request, and processing the extracted enterprise sales data based on the data analysis algorithm and enterprise sales data keywords, specifically includes: Determine the enterprise sales data keywords based on the preset keyword table; Determining a data arrangement type based on the enterprise sales data keywords, so as to group and arrange the enterprise sales data according to the data arrangement type; Based on the query request, determine a data analysis type; wherein the data analysis type includes at least one of data summation, data comparison, and data grouping; A data analysis algorithm is determined based on the data analysis type, so as to process the grouped and arranged enterprise sales data based on the data analysis algorithm.
5. The method for quickly extracting enterprise sales data according to claim 1, characterized in that: The form conversion of the processed enterprise sales data specifically includes: The received enterprise sales data is converted into a natural language form or a visual report form; wherein the received enterprise sales data is structured data.
6. The method for quickly extracting enterprise sales data according to claim 1, characterized in that: After receiving the query request sent by the user, the method further includes: Matching is performed in a history database based on the query request to obtain historical query text information; wherein the historical query text information includes at least one of the historical question record, the department to which the user belongs, and the current sales business scenario; Based on the historical query text information, the query request and the first preset deep learning model, the query request is expanded to obtain multi-dimensional query data; Furthermore, based on the historical query text information, the query request and the second preset deep learning model, data analysis task recommendations and self-service analysis operation task recommendations are performed on the query request.
7. A method for rapidly extracting enterprise sales data according to claim 6, characterized in that: After matching in the history database based on the query request to obtain historical query text information, the method further includes: When the query request meets the preset analysis difficulty condition, determining a corresponding machine learning model; Based on the historical query text information corresponding to the query request, the machine learning model is trained to obtain the required machine model; The query request is analyzed based on the required machine model.
8. The method for quickly extracting enterprise sales data according to claim 1, characterized in that: After classifying and storing the enterprise sales data obtained from different data sources, the method further includes: Identify sensitive data of the stored enterprise sales data, encrypt the identified sensitive data, and encrypt and transmit the enterprise sales data through the SSL / TLS encryption protocol; Based on data types, the encrypted enterprise sales data is divided; wherein the data types include at least one of role data types, department data types, and business demand data types; Based on the data partitioning results, different data access permissions are assigned to users.
9. A device for quickly extracting enterprise sales data, characterized in that: The device comprises a memory for storing computer program instructions and a processor for executing the program instructions, wherein when the computer program instructions are executed by the processor, the device is triggered to execute the method according to any one of claims 1 to 8.
10. A non-volatile computer storage medium storing computer executable instructions, characterized in that: The computer executable instructions can execute the method according to any one of claims 1 to 8.