Table data processing method and device and storage medium

Abstract

The invention relates to a table data processing method and device and a storage medium. The method comprises the steps of selecting to-be-analyzed target data from a table; acquiring historical data information of the user; and obtaining a plurality of insight themes in combination with the historical data information and the target data, and generating a corresponding visual chart for each insight theme. Therefore, the efficiency and accuracy of data analysis can be improved, a structured and easily understood data display mode is provided for the user, and meanwhile, the analysis result can better meet the requirements and preferences of the user, so that clear and professional support can be provided for decision making of the user, and the depth and effect of analysis are improved.

Description

Technical Field

[0001] This application relates to the field of computers, and in particular, to a method, apparatus, and storage medium for processing tabular data. Background Art

[0002] In the fields of data analysis and presentation, the chart function, as a basic ability of tables, is widely used among user groups in all walks of life, especially in key areas such as enterprise decision-making, data reporting, and business analysis. Charts can not only help users intuitively understand data but also reveal the trends and correlations behind the data, providing strong support for decision-making.

[0003] Current chart automatic generation technologies mainly rely on rules of data types, dimensions, and trends to understand the semantic structure of queries and accordingly automatically match suitable chart types. For example, when the data is numerical and arranged in chronological order, the system will recommend a line chart to show the trend of data over time; when the data contains categories, bar charts, stacked charts, or pie charts become recommended options for clearly showing the data distribution and comparison of each category; and when the data involves the relationship between multiple variables, scatter plots or heatmaps, etc., are regarded as ideal choices for revealing the potential associations and distribution patterns between variables.

[0004] These technologies have improved the degree of automation of chart generation to a certain extent, but there is still much room for improvement in the accuracy of charts and personalized recommendations for specific users. Summary of the Invention

[0005] This application provides a method, apparatus, and storage medium for processing tabular data to solve the technical problems that the existing chart automatic generation technologies have poor accuracy and cannot meet the personalized needs of users.

[0006] In a first aspect, this application provides a method for processing tabular data, the method including:

[0007] Select target data to be analyzed from the table;

[0008] Obtain the historical data information of the user;

[0009] Combine the historical data information and the target data to obtain multiple insight themes, and generate corresponding visualization charts for each of the insight themes.

[0010] In a possible implementation manner, the combining the historical data information and the target data to obtain multiple insight themes, and generating corresponding visualization charts for each of the insight themes includes:

[0011] Input the historical data information and the target data into a first insight model to obtain multiple insight themes;

[0012] Input the multiple insight topics and the target data into a second insight model to obtain a visualization chart corresponding to each insight topic;

[0013] Among them, inputting the historical data information and the target data into the first insight model to obtain multiple insight topics includes:

[0014] Structurally process the target data according to a set structural processing method to obtain the structured data corresponding to the target data;

[0015] Input the historical data information and the structured data corresponding to the target data into the first insight model to obtain multiple insight topics;

[0016] The step of inputting the multiple insight topics and the target data into the second insight model to obtain a visualization chart corresponding to each insight topic includes:

[0017] Input the multiple insight topics and the structured data corresponding to the target data into the second insight model to obtain a visualization chart corresponding to each insight topic.

[0018] In a possible implementation manner, after generating a corresponding visualization chart for each insight topic, it further includes:

[0019] Integrate and process the multiple insight topics and their corresponding visualization charts to form a data analysis report for the target data;

[0020] Among them, integrating and processing the multiple insight topics and their corresponding visualization charts to form a data analysis report for the target data includes:

[0021] Quantitatively score each insight topic to obtain a quantitative scoring result for each insight topic;

[0022] According to the quantitative scoring results, screen out multiple target insight topics from the multiple insight topics;

[0023] Integrate and process the multiple target insight topics and their corresponding visualization charts to form a data analysis report for the target data.

[0024] In a possible implementation manner, the step of quantitatively scoring each insight topic to obtain a quantitative scoring result for each insight topic includes:

[0025] Perform the following processing for each insight topic:

[0026] Quantify and score the insight theme using different quantification scoring methods to obtain sub-scoring results of the insight theme corresponding to each quantification scoring method;

[0027] Determine the quantification scoring result of the insight theme according to the sub-scoring results of the insight theme corresponding to each quantification scoring method.

[0028] In a possible implementation, the step of quantifying and scoring the insight theme using different quantification scoring methods to obtain sub-scoring results of the insight theme corresponding to each quantification scoring method includes:

[0029] Quantify and score the insight theme using a first quantification scoring method to obtain a sub-scoring result of the insight theme corresponding to the first quantification scoring method; and, quantify and score the insight theme using a second quantification scoring method to obtain a sub-scoring result of the insight theme corresponding to the second quantification scoring method;

[0030] Among them, the first quantification scoring method scores the insight theme based on user intent. The step of quantifying and scoring the insight theme using the first quantification scoring method includes:

[0031] Perform intent recognition on the historical data information of the user to obtain the intent of the user; determine the relevance between the insight theme and the intent of the user; determine the sub-scoring result of the insight theme corresponding to the first quantification scoring method according to the relevance;

[0032] The second quantification scoring method scores the insight theme based on table semantics. The step of quantifying and scoring the insight theme using the second quantification scoring method includes: determining the associated fields of the insight theme in the table; determining the semantic relevance between the insight theme and its associated fields; determining the sub-scoring result of the insight theme corresponding to the second quantification scoring method according to the semantic relevance.

[0033] In a possible implementation, the step of determining the quantification scoring result of the insight theme according to the sub-scoring results of the insight theme corresponding to each quantification scoring method includes:

[0034] Perform weighted summation processing on the sub-scoring results of the insight theme corresponding to each quantification scoring method to obtain the quantification scoring result of the insight theme.

[0035] In a possible implementation, the step of integrating and processing multiple target insight themes and their corresponding visualization charts to form a data analysis report for the target data includes:

[0036] Perform clustering processing on multiple target insight themes to obtain different insight theme clusters;

[0037] For each of the insight theme clusters, generate high-level insight themes for the insight theme clusters;

[0038] Typeset the target insight themes belonging to the same insight theme cluster and their corresponding visualization charts to obtain a summary page for each of the insight theme clusters, where the summary page includes the high-level insight theme corresponding to the insight theme cluster;

[0039] Typeset the summary pages of different insight theme clusters to form a data analysis report for the target data.

[0040] In a possible implementation manner, the clustering the multiple target insight themes to obtain different insight theme clusters includes:

[0041] Construct semantic vectors for each of the target insight themes; based on the semantic vectors of the target insight themes, cluster the multiple target insight themes to obtain different insight theme clusters;

[0042] Alternatively, based on the chart types of the visualization charts corresponding to the target insight themes, cluster the multiple target insight themes to obtain different insight theme clusters, where the visualization charts corresponding to the target insight themes in the same cluster have the same chart type, and the visualization charts corresponding to the target insight themes in different clusters have different chart types.

[0043] In a possible implementation manner, the selecting the target data to be analyzed from the table includes:

[0044] Select the target data to be analyzed from two or more tables respectively;

[0045] The obtaining multiple insight themes by combining the historical data information and the target data includes:

[0046] Perform the following processing on the target data in each of the tables: combine the historical data information and the target data in the table to obtain multiple insight themes;

[0047] And / or, take the target data in the two or more tables as a whole, combine the historical data information and the target data to obtain multiple insight themes.

[0048] In a possible implementation manner, the method further includes:

[0049] Generate chart insight information for the visualization charts of each of the insight themes, and integrate the chart insight information in the data analysis report of the target data.

[0050] In a second aspect, the present application provides an apparatus for processing tabular data, the apparatus comprising:

[0051] A target data selection module, configured to select target data to be analyzed from a table;

[0052] A historical data acquisition module, configured to acquire historical data information of a user;

[0053] An insight module, configured to combine the historical data information and the target data to obtain a plurality of insight themes, and generate corresponding visualization charts for each of the insight themes.

[0054] In a possible implementation manner, the insight module includes:

[0055] A theme acquisition unit, configured to input the historical data information and the target data into a first insight model to obtain a plurality of insight themes;

[0056] A chart generation unit, configured to input the plurality of insight themes and the target data into a second insight model to obtain visualization charts corresponding to each of the insight themes;

[0057] Wherein, the theme acquisition unit is specifically configured to:

[0058] Perform structured processing on the target data according to a set structured processing method to obtain structured data corresponding to the target data;

[0059] Input the historical data information and the structured data corresponding to the target data into a first insight model to obtain a plurality of insight themes;

[0060] The chart generation unit is specifically configured to:

[0061] Input the plurality of insight themes and the structured data corresponding to the target data into a second insight model to obtain visualization charts corresponding to each of the insight themes.

[0062] In a possible implementation manner, the apparatus further includes: an integration module;

[0063] The integration module includes:

[0064] A scoring unit, configured to perform quantitative scoring on each of the insight themes respectively to obtain a quantitative scoring result for each of the insight themes;

[0065] A screening unit, configured to screen out a plurality of target insight themes from the plurality of insight themes according to the quantitative scoring result;

[0066] An integration processing unit for integrally processing the multiple target insight topics and their corresponding visualization charts to form a data analysis report for the target data.

[0067] In a possible implementation manner, the scoring unit is specifically configured to:

[0068] Perform the following processing for each of the insight topics:

[0069] Quantitatively score the insight topic using different quantitative scoring methods to obtain sub-scoring results of the insight topic corresponding to each quantitative scoring method;

[0070] Determine the quantitative scoring result of the insight topic according to the sub-scoring results of the insight topic corresponding to each quantitative scoring method.

[0071] In a possible implementation manner, the scoring unit includes:

[0072] A first sub-scoring unit for quantitatively scoring the insight topic using a first quantitative scoring method to obtain a sub-scoring result of the insight topic corresponding to the first quantitative scoring method;

[0073] A second sub-scoring unit for quantitatively scoring the insight topic using a second quantitative scoring method to obtain a sub-scoring result of the insight topic corresponding to the second quantitative scoring method;

[0074] Wherein, the first quantitative scoring method scores the insight topic based on the user's intention, and the first sub-scoring unit is specifically configured to:

[0075] Perform intention recognition on the historical data information of the user to obtain the intention of the user; determine the relevance between the insight topic and the intention of the user; determine the sub-scoring result of the insight topic corresponding to the first quantitative scoring method according to the relevance;

[0076] The second quantitative scoring method scores the insight topic based on the table semantics, and the second sub-scoring unit is specifically configured to: determine the associated field of the insight topic in the table; determine the semantic relevance between the insight topic and its associated field; determine the sub-scoring result of the insight topic corresponding to the second quantitative scoring method according to the semantic relevance.

[0077] In a possible implementation manner, the scoring unit determines the quantitative scoring result of the insight topic according to the sub-scoring results of the insight topic corresponding to each quantitative scoring method, including:

[0078] Perform weighted summation processing on the sub-scoring results of the insight topic corresponding to each quantitative scoring method to obtain the quantitative scoring result of the insight topic.

[0079] In a possible implementation, the integration processing unit includes:

[0080] A clustering subunit, configured to perform clustering processing on multiple target insight themes to obtain different insight theme clusters;

[0081] A high-level theme extraction subunit, configured to generate a high-level insight theme for each insight theme cluster;

[0082] A first layout subunit, configured to perform layout processing on the target insight themes belonging to the same insight theme cluster and their corresponding visualization charts to obtain a summary page for each insight theme cluster, where the summary page includes the high-level insight theme corresponding to the insight theme cluster;

[0083] A second layout subunit, configured to perform layout processing on the summary pages of different insight theme clusters to form a data analysis report for the target data.

[0084] In a possible implementation, the clustering subunit is specifically configured to:

[0085] Construct a semantic vector for each target insight theme; based on the semantic vector of the target insight theme, perform clustering processing on multiple target insight themes to obtain different insight theme clusters;

[0086] Alternatively, based on the chart type of the visualization chart corresponding to the target insight theme, perform clustering processing on multiple target insight themes to obtain different insight theme clusters, where the chart types of the visualization charts corresponding to the target insight themes in the same cluster are the same, and the chart types of the visualization charts corresponding to the target insight themes in different clusters are different.

[0087] In a possible implementation, the target data selection module is specifically configured to:

[0088] Select target data to be analyzed from two or more tables respectively;

[0089] The insight module is specifically configured to:

[0090] Perform the following processing on the target data in each table: combine the historical data information and the target data in the table to obtain multiple insight themes;

[0091] And / or, take the target data in the two or more tables as a whole, combine the historical data information and the target data to obtain multiple insight themes.

[0092] In a possible implementation, the device further includes:

[0093] A summary module, configured to generate chart insight information for the visualization charts of each of the said insight topics, and integrate the said chart insight information into the data analysis report of the said target data.

[0094] In a third aspect, the present application provides an electronic device, including: a processor and a memory, where the processor is configured to execute a processing program for tabular data stored in the memory to implement the tabular data processing method according to any one of the first aspects.

[0095] In a fourth aspect, the present application provides a storage medium storing one or more programs, where the one or more programs can be executed by one or more processors to implement the tabular data processing method according to any one of the first aspects.

[0096] The above technical solutions provided by the embodiments of the present application have the following advantages compared with the prior art: The method provided by the embodiments of the present application, by obtaining the historical data information of the user, combining the historical data information and the target data to be analyzed, obtains multiple insight topics, and generates corresponding visualization charts for each insight topic, provides a method of combining the characteristics of the user's historical data information and the target data, automatically extracting valuable insight topics, and generating intuitive visualization charts, which can not only improve the efficiency and accuracy of data analysis, provide a structured and easy-to-understand data display method for users, but also make the analysis results more in line with the needs and preferences of users, so as to provide clear and professional support for user decision-making and enhance the depth and effect of analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0097] The drawings here are incorporated into the specification and constitute a part of this specification, showing embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application.

[0098] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0099] One or more embodiments are illustrated by way of example in the pictures in the corresponding drawings. These exemplary illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are represented as similar elements, unless otherwise stated, and the drawings in the drawings do not constitute a proportional limitation.

[0100] Figure 1 It is a flowchart of an embodiment of a method for processing tabular data provided by an embodiment of the present application;

[0101] Figure 2 This is a flowchart of an embodiment of another method for processing tabular data provided by an embodiment of the present application;

[0102] Figure 3 This is a flowchart of an embodiment of another method for processing tabular data provided by an embodiment of the present application;

[0103] Figure 4 This is a schematic diagram of a user selecting target data to be analyzed in a table;

[0104] Figure 5 This is a schematic diagram of an interface when presenting a data analysis report;

[0105] Figure 6 This is a schematic diagram of an interface when presenting a data analysis report;

[0106] Figure 7 This is a block diagram of an embodiment of an apparatus for processing tabular data provided by an embodiment of the present application;

[0107] Figure 8 This is a schematic diagram of the structure of an electronic device provided by an embodiment of the present application. Detailed implementation manners

[0108] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Apparently, the described embodiments are some, but not all, of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0109] The following disclosure provides many different embodiments or examples for implementing different structures of the present application. To simplify the disclosure of the present application, components and settings of specific examples are described below. Of course, they are only examples and are not intended to limit the present application. In addition, the present application may repeat reference numerals and / or letters in different examples. Such repetition is for the purpose of simplification and clarity and does not in itself indicate the relationship between the various embodiments and / or settings discussed.

[0110] To solve the technical problems of poor accuracy in the existing chart automatic generation technology and inability to meet the personalized needs of users, the present application provides a method, device and storage medium for processing tabular data, which can improve the efficiency and accuracy of data analysis, provide a structured and easy-to-understand data display method for users, and at the same time make the analysis results more in line with the needs and preferences of users, so as to provide clear and professional support for user decision-making and enhance the depth and effect of analysis.

[0111] Figure 1 This is a flowchart of an embodiment of a method for processing tabular data provided by an embodiment of the present application. As Figure 1 shown, the method includes the following steps:

[0112] Step 101: Select target data to be analyzed from the table.

[0113] A table usually consists of fields and data items. Among them, fields are column headers or header rows used to describe data attributes, which provide context and meaning for the data, such as "product name", "product category", "total sales amount", etc. Data items are specific data values filled below the fields, which represent actual information or numerical values, such as the specific name, specific category, and specific total sales amount of a certain product.

[0114] Analyzing the data in the table means conducting in-depth research and understanding on some or all of the fields in the table and the data items corresponding to these fields. This analysis not only involves data items (i.e., specific numerical values or information), but also includes the fields themselves (i.e., the attributes and dimensions of the data). Therefore, the target data to be analyzed includes both data items and fields. Among them, fields are the basis of analysis, defining the structure and dimensions of the data, while data items provide specific information.

[0115] As an optional implementation manner, the user is allowed to manually select the target data to be analyzed in the table. Specifically, in response to the data range selection operation performed by the user on the table, the data selected by this selection operation is determined as the target data to be analyzed. Exemplarily, the data range selection operation can be an intuitive selection operation, that is, the user directly selects the data area in the table by clicking and dragging, or the user can specify the required data area by inputting the row and column ranges.

[0116] As another alternative implementation, it is allowed to automatically select the target data to be analyzed from a table. Specifically, based on data analysis and machine learning techniques, the keyword fields and data items in the table can be automatically identified, and the target data to be analyzed can be intelligently filtered according to the user's analysis purpose and requirements. In this way, the user can automatically provide accurate and relevant data support without manually selecting the data, thus greatly improving the efficiency and accuracy of data analysis.

[0117] As an alternative implementation, the target data to be analyzed selected from the table can come from one table or multiple tables. When the target data comes from multiple tables, when the user manually selects the target data to be analyzed in the table, the target data to be analyzed can be selected from different tables successively, and the data from different tables can be listed in turn in Figure 4 the option box shown, or multiple option boxes can be added based on the quantity requirements of the selected tables, and the data from different tables can be listed separately in the corresponding option boxes. When the target data comes from multiple tables, when automatically selecting the target data to be analyzed from the table, based on data analysis and machine learning techniques, the keyword fields and data items in multiple tables can be automatically identified, and according to the user's analysis purpose and requirements, as well as the association relationships between multiple tables and the association relationships between the keyword fields and data items in multiple tables, the target data to be analyzed can be intelligently filtered.

[0118] Step 102: Obtain the user's historical data information.

[0119] Step 103: Combine the historical data information and the target data to obtain multiple insight themes, and generate corresponding visual charts for each insight theme.

[0120] Exemplarily, the insight themes may involve aspects such as data trends, associations, anomalies, etc. The visual charts may include various chart types such as line charts, bar charts, pie charts, scatter plots, etc., which specifically depend on the characteristics of the data and the requirements of the analysis. Those skilled in the art can understand that generating corresponding visual charts for each insight theme can help the user more intuitively understand the information and rules behind the data.

[0121] In one embodiment, the historical data information of the user includes the user's historical table data processing information, which can directly reflect the user's past habits, concerns, commonly used analysis methods or tools when processing table data, as well as the final generated data analysis reports or visualization types, etc. These information help the execution entity of the embodiment of the present application predict the possible needs and preferences of the user in the current table processing scenario. Then, by combining these information to extract insight themes from the target data and generating corresponding visualization charts for each insight theme, the accuracy of the extracted insight themes and the generated corresponding visualization charts can be significantly improved.

[0122] Among them, the user's historical table data processing information includes the user's historical table data processing information of the current table (referring to the table involved in step 101) or the associated table of the current table, and can also include the historical table data processing information of different users for the current table or the associated table of the current table. Here, when multiple users process data for the same table, their respective behavior records are also of reference value, because these records can reveal the analysis focuses and preferences of different user groups, providing the execution entity of the embodiment of the present application with multi-angle insights into user needs. Among them, the associated table of the current table can be a data table with a similar header structure to the current table, and this type of data table contains the same or similar fields as the current table. The associated table of the current table can also be a data table with a business logic relationship with the current table. For example, the associated table may be the pre-processing or post-processing result of the current table, or other data entities that participate in a certain business process together with the current table. The associated table of the current table can also be a data table that records the state or change trend of the current table in a past period of time, and so on. By analyzing the recommendation situation of the insight themes and / or charts of the associated table, it can provide auxiliary reference for the generation situation of the insight themes and / or charts of the target data.

[0123] As an optional implementation manner, the user's historical table data processing information is specifically manifested as the historical conversations between the user and the execution entity of the embodiment of the present application. These conversations are centered around the user's analysis needs. For example, the historical conversation is "Please help me analyze the change trend of sales in the past year and present it in the form of a line chart". Such historical conversations not only directly reflect the user's analysis purpose, but also reveal the data visualization methods preferred by the user, thus being able to provide a clear analysis orientation for the execution entity of the embodiment of the present application.

[0124] As another alternative implementation, the historical table data processing information of the user focuses on the user's actual operation behaviors, especially those related to manual chart generation. These behavior records can detail every step of the user's operations when processing table data, such as selecting the data range, setting the chart type, adjusting the chart style, etc. By deeply analyzing these operation behaviors, the execution entity of the embodiments of the present application can more meticulously understand the user's operation habits and analysis processes, thereby providing insight results that better meet the user's needs.

[0125] As another alternative implementation, the historical table data processing information of the user can also be the user's actual operation behaviors, which include the selection behaviors of historical users based on recommended insight topics and / or charts. The execution entity of the embodiments of the present application can more meticulously understand the user's historical operation habits and analysis processes, thereby providing insight results that better meet the user's needs.

[0126] In another embodiment, the historical data information of the user includes the user type to which the user belongs. In this embodiment, considering that for different types of users, their concerns and preferences usually vary, and these differences will directly affect the user's information needs and interpretation methods. Then, obtaining insight topics based on the user type can more accurately locate the user's needs and provide personalized analysis. In this embodiment, as an alternative implementation, the insight topic is a topic generated based on the historical data information, that is, the user type, rather than obtained from the table where the target data to be analyzed is located.

[0127] Exemplarily, the historical data information can be manually input by the user, automatically uploaded by the user, or automatically obtained based on the acquisition scope granted by the user. The embodiments of the present application do not limit this.

[0128] The technical solution provided by the embodiments of the present application, by obtaining the historical data information of the user, combining the historical data information and the target data to be analyzed in the table, obtaining multiple insight topics, and generating corresponding visual charts for each insight topic, provides a method for automatically extracting valuable insight topics and generating intuitive visual charts by combining the characteristics of the user's historical data information and the target data. This not only improves the efficiency and accuracy of data analysis, provides a structured and easy-to-understand data display method for users, but also makes the analysis results more in line with the user's needs and preferences, thereby being able to provide clear and professional support for user decision-making and enhancing the depth and effect of the analysis.

[0129] In one embodiment, after generating corresponding visualization charts for each insight theme, the multiple insight themes and their corresponding visualization charts are further integrated to form a data analysis report for the target data. Thus, the automatic generation of the data analysis report is achieved, and this data analysis report can clearly display each insight theme, helping users to more deeply understand the internal laws and trends of the target data.

[0130] Among them, as an optional implementation method, the multiple insight themes and their corresponding visualization charts can be integrated according to a certain organizational structure. For example, the multiple insight themes can be sorted and / or grouped according to aspects such as the relevance and importance of the insight themes. Next, based on the sorting and grouping results, these insight themes are embedded into the data analysis report template to form a complete data analysis report. Among them, the data analysis report template is a preset standardized template with a rigorous structure and a unified format, which is convenient for quickly generating professional reports. It can also be a personalized template arranged by the user according to personal preferences, report styles, or specific requirements. This method allows users to create a personalized data analysis report according to their actual situations and aesthetic preferences, improving the user experience.

[0131] In addition, in one embodiment, chart insight information can also be generated for the visualization charts of each insight theme and integrated into the data analysis report of the target data. That is to say, the finally formed data analysis report can include both the detailed information of each insight theme and the corresponding chart display, as well as some summary analyses or suggestions, thus being able to help users better understand the data and make decisions.

[0132] In one embodiment, in step 103 above, an exemplary implementation of obtaining multiple insight themes by combining historical data information and target data and generating corresponding visualization charts for each insight theme based on the target data includes:

[0133] Input the historical data information and the target data into the first insight model to obtain multiple insight themes. Subsequently, input the multiple insight themes and the target data into the second insight model to obtain the visualization charts corresponding to each insight theme.

[0134] Among them, inputting the historical data information and the target data into the first insight model to obtain multiple insight themes includes: performing structured processing on the target data according to a set structured processing method to obtain the structured data corresponding to the target data; inputting the historical data information and the structured data corresponding to the target data into the first insight model to obtain multiple insight themes

[0135] Input multiple insight themes and target data into the second insight model to obtain a visualization chart corresponding to each insight theme, including: input the structured data corresponding to multiple insight themes and target data into the second insight model to obtain a visualization chart corresponding to each insight theme.

[0136] Among them, the first insight model and the second insight model can be the same large language model or different large language models.

[0137] Among them, according to the set structured processing method, the target data is structured to obtain the structured data corresponding to the target data, aiming to transform the target data into a form that can be efficiently recognized and processed by an insight model (such as a large language model).

[0138] As an optional implementation method, the process of structuring the target data according to the set structured processing method includes: First, split the target data into three core components: a header list, a row data list, and column data (presented in the form of a key-value pair list). Among them, the header list contains the title fields in the target data, providing clear identifiers for each column of data. For example, the header list contains column headings such as product requirements, person in charge, R & D plan, progress, completion date, etc. The row data list contains the specific data values corresponding to each header, presented in rows. For example, a row of data may be represented as "[requirement 1, person in charge a, 2023 / 4 / 7, 70%, 2023 / 8 / 10]", corresponding to the values of fields such as product requirements, person in charge, R & D plan start date, progress percentage, and completion date respectively.

[0139] After completing the structuring of the target data, input the obtained structured data together with the historical data information into the insight model. The insight model deeply analyzes these data, extracts multiple insightful insight themes from them, and generates corresponding visualization charts for each insight theme. These charts not only intuitively display the characteristics and trends of the data but also provide valuable reference information for decision-makers, helping them make more informed decisions.

[0140] Through the above embodiments, the conversion process from the original tabular data to insight themes and visualization charts is realized, which can improve the efficiency and accuracy of data analysis.

[0141] In another embodiment, in step 103 above, when the target data to be analyzed selected from the table is included in multiple tables, combining the historical data information and the target data to obtain multiple insight themes includes:

[0142] Obtain the table description corresponding to the target data;

[0143] Obtain multiple insight themes by combining historical data information, target data, and the table descriptions corresponding to the target data.

[0144] Among them, the table description corresponding to the target data is the association relationship between multiple tables included in the target data and the association relationship between each field in multiple tables included in the target data; as an optional implementation, the association relationship between each field in multiple tables can be the association relationship between each field in multiple tables or the association relationship between each field in each table included in multiple tables.

[0145] Among them, as an optional implementation, when the target data to be analyzed selected from the table is included in multiple tables, combine the historical data information and the target data to obtain multiple insight themes, including:

[0146] Input the target data into the third insight model to obtain the table description corresponding to the target data;

[0147] Input the historical data information, target data, and the table description corresponding to the target data into the first insight model to obtain multiple insight themes. Among them, the third insight model and the first insight model can be the same large language model or different large language models; it can be the same large language model or two large language models can be used.

[0148] Or, as an optional implementation, when the target data to be analyzed selected from the table is included in multiple tables, combine the historical data information and the target data to obtain multiple insight themes, including:

[0149] Input the historical data information and the target data into the fourth insight model to obtain the table description corresponding to the target data;

[0150] Input the historical data information, target data, and the table description corresponding to the target data into the first insight model to obtain multiple insight themes. Among them, the fourth insight model and the first insight model can be the same large language model or different large language models; it can be the same large language model or two large language models can be used.

[0151] Among them, as an optional embodiment, input the historical data information, target data, and the table description corresponding to the target data into the first insight model to obtain multiple insight themes, including:

[0152] According to the set structured processing method, combine the table description corresponding to the target data to perform structured processing on the target data to obtain the structured data corresponding to the target data;

[0153] Input the historical data information and the structured data corresponding to the target data into the first insight model to obtain multiple insight themes.

[0154] Among them, as an optional embodiment, when inputting the historical data information and the target data into the fourth insight model together, corresponding multiple tables and the fields included in the multiple tables can be determined in combination with the content included in the historical data information, and table descriptions can be generated in a targeted manner, so that table descriptions corresponding to the target data can be obtained more accurately and efficiently based on the user's purpose and / or historical habits.

[0155] Furthermore, when the target data to be analyzed selected from the tables is included in multiple tables, multiple insight themes are obtained by combining the historical data information and the target data, and corresponding visual charts are generated for each of the insight themes, including:

[0156] Input the target data into the third insight model to obtain the table description corresponding to the target data;

[0157] Input the historical data information and the target data into the first insight model to obtain multiple insight themes;

[0158] Input the multiple insight themes and the target data into the second insight model to obtain the visual chart corresponding to each of the insight themes.

[0159] Through the above embodiments, valuable insight themes are effectively extracted from multiple tables, providing strong support for data analysis and decision-making.

[0160] Figure 2 It is a flowchart of an embodiment of another method for processing tabular data provided by an embodiment of this application. Figure 2 The process shown is Figure 1 Based on the process shown, it focuses on describing how to integrally process multiple insight themes and their corresponding visual charts to form a data analysis report for the target data. As Figure 2 shown, it includes the following steps:

[0161] Step 201: Select the target data to be analyzed from the tables.

[0162] Step 202: Obtain the historical data information of the user.

[0163] Step 203: Combine the historical data information and the target data to obtain multiple insight themes, and generate corresponding visual charts for each insight theme.

[0164] For the detailed descriptions of steps 201 to 203, please refer to the relevant descriptions in the process shown in Figure 1 and will not be elaborated here.

[0165] Step 204: Perform a quantitative scoring for each insight theme respectively to obtain the quantitative scoring results of each insight theme.

[0166] Step 205: According to the quantitative scoring results of each insight theme, screen out multiple target insight themes from multiple insight themes.

[0167] Step 206: Integrate and process multiple target insight themes and their corresponding visualization charts to form a data analysis report for the target data.

[0168] It can be seen from the descriptions of steps 204 to 206 that in the process of generating the data analysis report, not all the insight themes extracted by the insight model are simply adopted, but a more refined and targeted strategy is adopted. Specifically, in step 204, a quantitative scoring is performed on each insight theme extracted by the insight model to measure the importance and influence of each insight theme through the quantitative scoring results, so as to provide an objective and quantitative basis for the subsequent screening work. Here, the quantitative scoring may be based on multiple factors, such as the uniqueness of the insight theme, the relevance to the business objective, the strength of data support, etc., and the embodiments of the present application do not limit this.

[0169] Subsequently, in step 205, according to the quantitative scoring results, multiple target insight themes are screened out from the multiple insight themes extracted by the insight model, aiming to ensure that the insight themes finally included in the data analysis report are the most valuable and can most attract the attention of users. Through this step, it can effectively avoid the overload of the finally generated data analysis report, making the report more focused, easier to understand and digest.

[0170] Finally, in step 206, the screened target insight themes and their corresponding visualization charts are integrated and processed, aiming to orderly and logically combine each target insight theme and visualization chart together to form a data analysis report with a clear structure and coherent content. Through the integration and processing, the integrity and readability of the finally generated data analysis report are strengthened, and the value of it as a decision-making support tool is further improved.

[0171] In summary, Figure 2 the process shown Figure 1 further optimizes the content and quality of the finally generated data analysis report on the basis of the effects that the process shown can achieve, so as to further improve the user experience and decision-making efficiency.

[0172] In one embodiment, the exemplary implementation of obtaining the quantitative scoring result of each insight theme by performing quantitative scoring on each insight theme separately includes: performing the following processing for each insight theme: performing quantitative scoring on the insight theme using different quantitative scoring methods to obtain sub-scoring results of the insight theme corresponding to each quantitative scoring method; and determining the quantitative scoring result of the insight theme according to the sub-scoring results of the insight theme corresponding to each quantitative scoring method.

[0173] Among them, as an optional implementation, the above different quantitative scoring methods include a first quantitative scoring method and a second quantitative scoring method. These two quantitative scoring methods perform quantitative scoring on the insight theme from different perspectives to obtain quantitative scoring results under two perspectives. Specifically, perform quantitative scoring on the insight theme using the first quantitative scoring method to obtain the sub-scoring result of the insight theme corresponding to the first quantitative scoring method; and perform quantitative scoring on the insight theme using the second quantitative scoring method to obtain the sub-scoring result of the insight theme corresponding to the second quantitative scoring method, and then jointly determine the quantitative scoring result of the insight theme according to the sub-scoring result of the insight theme corresponding to the first quantitative scoring method and the sub-scoring result corresponding to the second quantitative scoring method.

[0174] Exemplarily, perform a weighted sum processing on the sub-scoring results of the insight theme corresponding to each quantitative scoring method to obtain the quantitative scoring result of the insight theme. Among them, the weight of the quantitative scoring method, as an index to measure the importance of different quantitative scoring methods, has high flexibility and pertinence in setting. In practical applications, different weight values can be assigned to each quantitative scoring method according to actual needs to reflect their relative importance in the overall evaluation. The embodiments of the present application do not limit this.

[0175] The above embodiment provides a comprehensive quantitative scoring mechanism. This mechanism does not rely solely on a single scoring method, but combines multiple quantitative scoring methods and jointly acts on the evaluation process of the insight theme. In this way, it is possible to more comprehensively and objectively evaluate the value of each insight theme, and it is also possible to flexibly adjust the evaluation focus according to actual needs, providing more accurate and valuable reference information for decision-makers.

[0176] In one embodiment, the first quantitative scoring method focuses on performing quantitative scoring on the insight theme from the perspectives of user needs and business goals, aiming to capture the internal connection between the insight theme and the actual needs of users. Exemplarily, evaluate whether the insight theme is related to the business goal, evaluate whether the insight theme can help solve actual business problems, or provide valuable reference for decision-making, evaluate whether the insight theme is meaningful to the target audience, etc. For example, for the marketing team, insights about customer behavior may be more valuable; for the product development team, insights about user needs may be more valuable.

[0177] As an alternative implementation, the first quantitative scoring method scores the insight topic based on the user's intention. The specific implementation of quantitatively scoring the insight topic using the first quantitative scoring method includes: identifying the intention of the user from the historical data information of the user to obtain the user's intention; determining the relevance between the insight topic and the user's intention; and determining the sub-scoring result of the insight topic corresponding to the first quantitative scoring method according to the relevance.

[0178] Among them, identifying the intention of the user from the historical data information of the user to obtain the user's intention aims to identify the insight direction that the user may be concerned about. Subsequently, the identified user intention is matched with multiple insight topics one by one to calculate the relevance between them, aiming to quantitatively evaluate the internal connection between the two. Finally, according to the calculation result of the relevance, a sub-scoring result is assigned to the insight topic. Among them, the higher the relevance, the higher the sub-scoring result obtained by the insight topic, which means that the insight topic is more meaningful and valuable to the user's needs and business goals.

[0179] Thus, the first quantitative scoring method can objectively evaluate the significance and value of the insight topic to the user and business goals, thereby providing valuable reference for decision-makers.

[0180] In one embodiment, the second quantitative scoring method focuses on quantitatively scoring the insight topic from the perspectives of data quality and insight innovation, aiming to ensure that the insight topic is based on a solid and reliable data foundation and can provide novel and valuable insights. Exemplarily, in evaluating data quality, it can be concerned whether the conclusion of the insight is based on reliable data and whether it can be fully supported by the original data. Among them, in the case where the original data contains target data obtained from multiple tables, in addition to paying attention to the data in a single table on which the conclusion of the insight is based, if the conclusion of the insight can be obtained based on the association relationship between multiple tables, this also means that the conclusion of the insight can be supported by the original data. It can also be concerned whether the reasoning process of the insight is reasonable and check whether the logical chain from data to conclusion is clear and rigorous, etc. In evaluating insight innovation, it can be concerned whether the insight provides new information or perspectives, and evaluate whether the insight merely repeats known conclusions or reveals patterns or trends that have not been noticed before.

[0181] As an alternative implementation, the second quantitative scoring method scores the insight topic based on table semantics. The specific implementation of quantitatively scoring the insight topic using the second quantitative scoring method includes: determining the associated fields of the insight topic in the table; determining the semantic relevance between the insight topic and its associated fields; and determining the sub-scoring result of the insight topic corresponding to the second quantitative scoring method according to the semantic relevance.

[0182] Among them, when the insight model extracts insight topics from target data, it can track and record the associated fields of the insight topics in the table. Here, the associated fields of the insight topics in the table are the data basis for the formation of the insight topics, and the associated fields directly participate in the formation of the insight topics. Among them, in the case of obtaining target data to be analyzed from multiple tables, the associated fields of the same insight topic can come from different tables. For example, the associated fields of the same insight topic include field A, field B, and field C, where field A comes from table 1, and fields B and C come from table 2. Field A in table 1, field B in table 2, and field C together form the data basis for this insight topic.

[0183] Subsequently, the semantic relevance between the insight topic and its associated fields is determined, aiming to quantify the internal connection between the insight topic and the original data, so as to ensure the reliability and accuracy of the insight conclusion. Finally, according to the calculation result of the semantic relevance, a corresponding sub-scoring result is assigned to the insight topic. This sub-scoring result will comprehensively reflect the performance of the insight topic in terms of data quality and innovation, providing a comprehensive and objective evaluation basis for decision-makers.

[0184] It can be seen that the second quantitative scoring method can objectively evaluate the data quality and reliability of the insight topic, as well as the innovation of the insight topic, thus providing valuable reference for decision-makers.

[0185] Figure 3 It is a flowchart of an embodiment of another processing method for table data provided by an embodiment of the present application. Figure 3 The process shown is Figure 1 and Figure 2 On the basis of the process shown, it describes how to integrate and process multiple target insight topics and their corresponding visualization charts to form a data analysis report for the target data. As Figure 3 shown, it includes the following steps:

[0186] Step 301: Perform clustering processing on multiple target insight topics to obtain different insight topic clusters.

[0187] In step 301, clustering processing is performed on multiple target insight topics to obtain different insight topic clusters, aiming to be able to integrate multiple insight topics and their corresponding visualization charts according to a certain organizational structure in the follow-up. Through this processing, similar or related insight topics can be grouped into one category, thus more clearly showing the internal structure and relevance of the data.

[0188] In one embodiment, an exemplary implementation of clustering multiple target insight topics to obtain different insight topic clusters includes: constructing a semantic vector for each target insight topic; and clustering the multiple target insight topics based on the semantic vectors of the target insight topics to obtain different insight topic clusters.

[0189] Among them, the construction of the semantic vector can be based on the text description of the target insight topic and can be implemented through a pre-trained language model (such as BERT, GPT, etc.) or a dedicated text embedding tool (such as Word2Vec, GloVe, etc.). After obtaining the semantic vector of the target insight topic, a clustering algorithm (such as K-means, DBSCAN, hierarchical clustering, etc.) can be used to cluster the target insight topics. The clustering algorithm will classify the target insight topics into different clusters according to the distance (or similarity) between the semantic vectors. The target insight topics in the same cluster are closer semantically, while the target insight topics in different clusters have greater semantic differences.

[0190] In another embodiment, an exemplary implementation of clustering multiple target insight topics to obtain different insight topic clusters includes: clustering the multiple target insight topics based on the chart types of the visualization charts corresponding to the target insight topics to obtain different insight topic clusters, where the visualization charts corresponding to the target insight topics in the same cluster have the same chart type, and the visualization charts corresponding to the target insight topics in different clusters have different chart types.

[0191] In this embodiment, the target insight topics in the same cluster will have the same chart type, which means they have similarities in data presentation methods and visualization requirements. While the target insight topics in different clusters have different chart types, reflecting the differences in their data characteristics and visualization requirements.

[0192] Step 302: For each insight topic cluster, generate a high-level insight topic for the insight topic cluster.

[0193] Step 303: Perform typesetting on the target insight topics belonging to the same insight topic cluster and their corresponding visualization charts to obtain a summary page for each insight topic cluster, where the summary page includes the high-level insight topic corresponding to the insight topic cluster.

[0194] Step 304: Perform typesetting on the summary pages of different insight topic clusters to form a data analysis report for the target data.

[0195] As can be seen from the above steps, the technical solution provided by the embodiments of the present application typesets the target insight topics belonging to the same insight topic cluster and their corresponding visualization charts, so as to generate a summary page for each insight topic cluster. This step effectively aggregates similar or related insight topics, making subsequent analysis and presentation more organized.

[0196] Among them, on the summary page of the insight topic cluster, in addition to each target insight topic and its corresponding visualization chart included in the cluster, it may also include the high-level insight topic corresponding to the insight topic cluster. The so-called high-level insight topic refers to the core content that can summarize all the target insight topics in the cluster, thus providing a way for users to quickly understand the content of the cluster. Exemplarily, topic modeling algorithms such as LDA (Latent Dirichlet Allocation) can be used to extract potential high-level insight topics from the insight topic cluster.

[0197] Among them, on the summary page of the insight topic cluster, multiple target insight topics and their corresponding visualization charts can be sorted and displayed based on factors such as the importance, relevance, and quantitative scoring results of the target insight topics, so as to further make subsequent analysis and presentation more organized.

[0198] Finally, the summary pages of different insight topic clusters are typeset as a whole to form a complete data analysis report. This step combines the summary pages of each cluster, making the entire report clearer and more coherent logically. Users can gradually understand and analyze the content of different insight topic clusters by reading the report, so as to comprehensively grasp the characteristics and trends of the data.

[0199] In addition, in practical applications, aesthetic adjustments can also be made to the entire data analysis report to ensure consistent styles and neat typesetting among all parts. This includes unified design of elements such as font size, line spacing, and page margins, as well as clear presentation of charts and text content. The embodiments of the present application do not limit this.

[0200] Figure 3 The shown process, through clustering multiple target insight topics to obtain different insight topic clusters, typesetting the target insight topics belonging to the same insight topic cluster and their corresponding visualization charts to obtain a summary page for each insight topic cluster, and typesetting the summary pages of different insight topic clusters to form a data analysis report for the target data, can obtain a data analysis report with clear organization and strong logic, so that users can obtain the required information faster, provide clear and professional support for user decision-making, and improve the depth and effect of analysis.

[0201] Based on the method for processing tabular data provided in the above embodiments, the following shows an exemplary application scenario of this method:

[0202] As Figure 4 shown, the user first selects the target data to be analyzed in the table. For example, the target data selected by the user includes the data range A1:K300 in the sales record table.

[0203] Subsequently, the user can trigger Figure 4 the "AI Generate Chart" button in the interface shown in Figure 5 . The execution subject of the embodiment of the present application responds to the trigger operation of this button, executes the method provided in the embodiment of the present application, generates an analysis report of the target data and displays it. As Figure 5 shown, it is a schematic diagram of the interface when displaying the data analysis report. Figure 5 The interface shown includes multiple insight themes, such as customer purchase frequency, combined purchase behavior, sales trend, promotion effect, return and after-sales, seasonal consumption, etc., as well as the corresponding visualization charts and chart insight information for each insight theme. The chart types of these visualization charts are not completely the same, and there are various types of charts such as pie charts, bar charts, and line charts, which specifically depend on the characteristics of the data and the requirements of the analysis.

[0204] Exemplarily, the execution subject of the embodiment of the present application also supports the user to view the chart details and detailed chart insight information, and supports the user to export the data analysis report. For example, referring to Figure 6 which is a schematic diagram of the interface when displaying the data analysis report, Figure 6 it shows the chart details and detailed chart insight information of the insight theme of "combined purchase behavior", so that the user can more clearly understand the insight details and improve the user experience.

[0205] When the user needs to export the data analysis report, they can click Figure 6 the "Export Data Report" button in the interface shown. The execution subject of the embodiment of the present application responds to the trigger operation of this button, calls the relevant interfaces of the electronic device, and exports the generated data analysis report to the electronic device or a storage medium external to the electronic device.

[0206] In addition, referring to Figure 5 which is Figure 5 shown, a user interaction Q&A entry is also provided on the interface, specifically "No suitable chart? Come to Data Q&A to state your needs". This means that the user can issue clear data analysis instructions through this function. For example, "Please help me analyze the sales trend in the past year and present it in the form of a line chart". At this time, the execution subject of the embodiment of the present application also supports generating insight themes and visualization charts that can meet the user's business goals for the user's data analysis instructions, so as to better meet the user's business goals.

[0207] As can be seen from the above exemplary application scenarios, the technical solution provided by the embodiments of the present application can significantly improve the efficiency and accuracy of data analysis through automated insight extraction and intelligent organization, avoid users from manually sorting and analyzing data, reduce the cognitive burden of users, and make the data analysis process more efficient and smooth. In addition, the technical solution provided by the embodiments of the present application can be seamlessly integrated into the human-machine dialogue process, enhance the depth and breadth of data exploration, help users more easily discover potential associations and make more accurate decisions.

[0208] In addition, in an exemplary application scenario, the technical solution provided by the embodiments of the present application can also be applied to perform data analysis on two or more tables. Specifically, in one embodiment,

[0209] The specific implementation of selecting the target data to be analyzed from the table includes: selecting the target data to be analyzed from two or more tables respectively. Correspondingly, the specific implementation of extracting multiple insight themes from the target data in combination with historical data information includes: for the target data in each table, obtaining multiple insight themes in combination with historical data information and the target data; and / or, taking the target data in two or more tables as a whole, and obtaining multiple insight themes in combination with historical data information and the target data in two or more tables.

[0210] Among them, the process of performing data analysis on the target data in different tables respectively can refer to the description of the above embodiments and will not be elaborated here.

[0211] When taking the target data in different tables as a whole and applying the technical solution provided by the embodiments of the present application for data analysis, it means considering not only the data characteristics within a single table, but also the interaction relationships between the data in different tables. At this time, the extracted insight themes may include not only trends, anomalies, etc. within a single table, but also can reveal the differences, connections, mutual influences or potential associations between the data in two or more tables. For example, through analysis, it may be found that there is a significant positive or negative correlation between a certain data indicator in Table A and another data indicator in Table B; or there is a corresponding explanation or corresponding change for a certain outlier in Table C in Table D.

[0212] It can be seen that the technical solution provided by the embodiments of the present application can effectively extract valuable insight themes from multiple tables, providing strong support for data analysis and decision-making.

[0213] Figure 7 It is a block diagram of an embodiment of a processing device for tabular data provided by an embodiment of the present application. As Figure 7 shown, the device includes:

[0214] A target data selection module 71, configured to select target data to be analyzed from a table;

[0215] A historical data acquisition module 72, configured to acquire historical data information of a user;

[0216] An insight module 73, configured to combine the historical data information and the target data to obtain a plurality of insight themes, and generate corresponding visualization charts for each of the insight themes.

[0217] In a possible implementation manner, the insight module 73 includes:

[0218] A theme acquisition unit, configured to input the historical data information and the target data into a first insight model to obtain a plurality of insight themes;

[0219] A chart generation unit, configured to input the plurality of insight themes and the target data into a second insight model to obtain visualization charts corresponding to each of the insight themes;

[0220] Wherein, the theme acquisition unit is specifically configured to:

[0221] Structurally process the target data according to a set structural processing method to obtain structured data corresponding to the target data;

[0222] Input the historical data information and the structured data corresponding to the target data into a first insight model to obtain a plurality of insight themes;

[0223] The chart generation unit is specifically configured to:

[0224] Input the plurality of insight themes and the structured data corresponding to the target data into a second insight model to obtain visualization charts corresponding to each of the insight themes.

[0225] In a possible implementation manner, the device further includes: an integration module;

[0226] The integration module includes:

[0227] A scoring unit, configured to perform quantitative scoring on each of the insight themes respectively to obtain a quantitative scoring result for each of the insight themes;

[0228] A screening unit, configured to screen out a plurality of target insight themes from the plurality of insight themes according to the quantitative scoring result;

[0229] An integration processing unit, configured to perform integration processing on the plurality of target insight themes and their corresponding visualization charts to form a data analysis report for the target data.

[0230] In a possible implementation, the scoring unit is specifically configured to:

[0231] Perform the following processing for each of the insight topics:

[0232] Quantitatively score the insight topic using different quantitative scoring methods to obtain sub-scoring results of the insight topic corresponding to each quantitative scoring method;

[0233] Determine the quantitative scoring result of the insight topic according to the sub-scoring results of the insight topic corresponding to each quantitative scoring method.

[0234] In a possible implementation, the scoring unit includes:

[0235] A first sub-scoring unit, configured to quantitatively score the insight topic using a first quantitative scoring method to obtain a sub-scoring result of the insight topic corresponding to the first quantitative scoring method;

[0236] A second sub-scoring unit, configured to quantitatively score the insight topic using a second quantitative scoring method to obtain a sub-scoring result of the insight topic corresponding to the second quantitative scoring method;

[0237] Wherein, the first quantitative scoring method scores the insight topic based on the user's intention, and the first sub-scoring unit is specifically configured to:

[0238] Perform intention recognition on the historical data information of the user to obtain the intention of the user; determine the relevance between the insight topic and the intention of the user; determine the sub-scoring result of the insight topic corresponding to the first quantitative scoring method according to the relevance;

[0239] The second quantitative scoring method scores the insight topic based on the table semantics, and the second sub-scoring unit is specifically configured to: determine the associated field of the insight topic in the table; determine the semantic relevance between the insight topic and its associated field; determine the sub-scoring result of the insight topic corresponding to the second quantitative scoring method according to the semantic relevance.

[0240] In a possible implementation, the scoring unit determines the quantitative scoring result of the insight topic according to the sub-scoring results of the insight topic corresponding to each quantitative scoring method, including:

[0241] Perform weighted summation processing on the sub-scoring results of the insight topic corresponding to each quantitative scoring method to obtain the quantitative scoring result of the insight topic.

[0242] In a possible implementation, the integration processing unit includes:

[0243] A clustering subunit, configured to perform clustering processing on multiple said target insight topics to obtain different insight topic clusters;

[0244] A high-level topic extraction subunit, configured to generate high-level insight topics for each said insight topic cluster;

[0245] A first layout subunit, configured to perform layout processing on the target insight topics belonging to the same insight topic cluster and their corresponding visualization charts to obtain a summary page for each said insight topic cluster, where the summary page includes the high-level insight topics corresponding to the insight topic cluster;

[0246] A second layout subunit, configured to perform layout processing on the summary pages of different insight topic clusters to form a data analysis report for the said target data.

[0247] In a possible implementation manner, the clustering subunit is specifically configured to:

[0248] Construct semantic vectors for each said target insight topic; based on the semantic vectors of the target insight topics, perform clustering processing on multiple said target insight topics to obtain different insight topic clusters;

[0249] Or, based on the chart types of the visualization charts corresponding to the target insight topics, perform clustering processing on multiple said target insight topics to obtain different insight topic clusters, where the chart types of the visualization charts corresponding to the target insight topics in the same cluster are the same, and the chart types of the visualization charts corresponding to the target insight topics in different clusters are different.

[0250] In a possible implementation manner, the target data selection module is specifically configured to:

[0251] Select target data to be analyzed from two or more tables respectively;

[0252] The insight module is specifically configured to:

[0253] Perform the following processing on the target data in each said table: combine the historical data information and the target data in the table to obtain multiple insight topics;

[0254] And / or, use the target data in the two or more tables as a whole, combine the historical data information and the target data to obtain multiple insight topics.

[0255] In a possible implementation manner, the apparatus further includes:

[0256] A summary module for generating chart insight information for the visualization charts of each of the said insight topics and integrating the said chart insight information into the data analysis report of the target data.

[0257] As Figure 8 shown, an embodiment of the present application provides an electronic device, including a processor 111, a communication interface 112, a memory 113, and a communication bus 114. Among them, the processor 111, the communication interface 112, and the memory 113 complete mutual communication through the communication bus 114.

[0258] The memory 113 is used to store computer programs.

[0259] In an embodiment of the present application, when the processor 111 executes the program stored on the memory 113, it implements the processing method of tabular data provided by any of the foregoing method embodiments, including:

[0260] Select target data to be analyzed from the table.

[0261] Obtain the historical data information of the user.

[0262] Combining the historical data information and the target data, obtain multiple insight topics and generate corresponding visualization charts for each of the said insight topics.

[0263] An embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the steps of the processing method of tabular data provided by any of the foregoing method embodiments.

[0264] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0265] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution or the part that contributes to the related technology can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0266] It should be understood that the terms used herein are for the purpose of describing particular example embodiments only and are not intended to be limiting. Unless the context clearly dictates otherwise, the singular forms "a", "an", and "the" as used herein may also include the plural forms. The terms "comprising", "including", "containing", and "having" are inclusive and thus specify the presence of stated features, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, steps, operations, elements, components, and / or combinations thereof. The method steps, processes, and operations described herein are not to be construed as necessarily requiring their performance in the particular order described or illustrated, unless an execution order is explicitly stated. It should also be understood that additional or alternative steps may be used.

[0267] The foregoing are only specific embodiments of the present application, which enable those skilled in the art to understand or implement the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but rather is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for processing tabular data, characterized in that: The method comprises: Select the target data to be analyzed from the table; Get the user's historical data information; Combine the historical data information and the target data to obtain multiple insight themes, and generate a corresponding visualization chart for each insight theme.

2. The method according to claim 1, characterized in that The combining of the historical data information and the target data to obtain a plurality of insight topics and generating a corresponding visualization chart for each insight topic includes: Inputting the historical data information and the target data into a first insight model to obtain a plurality of insight topics; Inputting the plurality of insight themes and the target data into a second insight model to obtain a visualization chart corresponding to each insight theme; The historical data information and the target data are input into the first insight model to obtain multiple insight topics, including: According to the set structured processing method, the target data is structured to obtain structured data corresponding to the target data; Inputting the historical data information and structured data corresponding to the target data into a first insight model to obtain a plurality of insight topics; Inputting the plurality of insight themes and the target data into a second insight model to obtain a visualization chart corresponding to each insight theme; including: The structured data corresponding to the multiple insight themes and the target data are input into a second insight model to obtain a visualization chart corresponding to each insight theme.

3. The method according to claim 1, characterized in that: After generating a corresponding visualization chart for each insight topic, the method further includes: Integrate the multiple insight themes and their corresponding visualization charts to form a data analysis report for the target data; The multiple insight themes and their corresponding visualization charts are integrated to form a data analysis report for the target data, including: Performing quantitative scoring on each of the insight topics respectively to obtain a quantitative scoring result for each of the insight topics; According to the quantitative scoring results, a plurality of target insight topics are screened out from the plurality of insight topics; The plurality of target insight themes and their corresponding visualization charts are integrated and processed to form a data analysis report for the target data.

4. The method according to claim 3, characterized in that The step of performing quantitative scoring on each insight theme to obtain a quantitative scoring result for each insight theme includes: The following processing is performed for each of the insight topics: Using different quantitative scoring methods to quantitatively score the insight theme, and obtaining a sub-scoring result of the insight theme corresponding to each quantitative scoring method; The quantitative scoring result of the insight theme is determined according to the sub-scoring result of each quantitative scoring method corresponding to the insight theme.

5. The method according to claim 4, characterized in that The quantitative scoring of the insight theme by using different quantitative scoring methods to obtain the sub-scoring results of the insight theme corresponding to each quantitative scoring method includes: Using a first quantitative scoring method to quantitatively score the insight theme, and obtaining a sub-scoring result of the insight theme corresponding to the first quantitative scoring method; and using a second quantitative scoring method to quantitatively score the insight theme, and obtaining a sub-scoring result of the insight theme corresponding to the second quantitative scoring method; The first quantitative scoring method scores the insight topic based on the user intention, and the quantitative scoring of the insight topic by using the first quantitative scoring method includes: Performing intention recognition on the historical data information of the user to obtain the intention of the user; determining the correlation between the insight topic and the intention of the user; and determining a sub-scoring result of the insight topic corresponding to the first quantitative scoring method according to the correlation; The second quantitative scoring method scores the insight theme based on the semantics of the table, and the quantitative scoring of the insight theme using the second quantitative scoring method includes: determining the associated fields of the insight theme in the table; determining the semantic relevance between the insight theme and its associated fields; and determining the sub-scoring result of the second quantitative scoring method corresponding to the insight theme based on the semantic relevance.

6. The method according to claim 4, characterized in that Determining the quantitative scoring result of the insight theme according to the sub-scoring result of each quantitative scoring method corresponding to the insight theme includes: The sub-scoring results of each quantitative scoring method corresponding to the insight theme are weighted and summed to obtain the quantitative scoring result of the insight theme.

7. The method according to claim 3, characterized in that The integrating and processing the plurality of target insight themes and their corresponding visualization charts to form a data analysis report for the target data includes: Clustering the plurality of target insight topics to obtain different insight topic clusters; For each of the insight theme clusters, generating a high-level insight theme for the insight theme cluster; Layout the target insight themes and their corresponding visualization charts belonging to the same insight theme cluster to obtain a summary page for each insight theme cluster, wherein the summary page contains high-level insight themes corresponding to the insight theme cluster; The summary pages of different insight theme clusters are typeset to form a data analysis report for the target data.

8. The method according to claim 7, characterized in that The clustering process of the plurality of target insight topics to obtain different insight topic clusters includes: Constructing a semantic vector for each of the target insight topics; clustering the multiple target insight topics based on the semantic vectors of the target insight topics to obtain different insight topic clusters; Alternatively, based on the chart type of the visualization chart corresponding to the target insight theme, multiple target insight themes are clustered to obtain different insight theme clusters, wherein the chart types of the visualization charts corresponding to the target insight themes in the same cluster are the same, and the chart types of the visualization charts corresponding to the target insight themes in different clusters are different.

9. The method according to claim 1, characterized in that: The step of selecting target data to be analyzed from the table includes: Select target data to be analyzed from two or more tables respectively; The combining of the historical data information and the target data to obtain multiple insight topics includes: The following processing is performed for each target data in the table: combining the historical data information and the target data in the table to obtain a plurality of insight topics; And / or, the target data in the two or more tables are taken as a whole, combined with the historical data information and the target data, to obtain multiple insight themes.

10. The method according to any one of claims 1 to 9, characterized in that: The method further comprises: Generate chart insight information for each visualization chart of the insight theme, and integrate the chart insight information into a data analysis report of the target data.

11. A device for processing tabular data, characterized in that: The device comprises: A target data selection module is used to select target data to be analyzed from a table; A historical data acquisition module is used to obtain historical data information of users; The insight module is used to combine the historical data information and the target data to obtain multiple insight themes and generate a corresponding visualization chart for each insight theme.

12. A storage medium, characterized in that: The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the method for processing table data according to any one of claims 1 to 10.

Images