A Method for Data Interactive Transformation and Visualization Embedding of Complex Tables

Abstract

The present invention belongs to the technical field of visualization and human-computer interaction. Specifically, it relates to a method for data interactive transformation and visualization embedding of complex tables, including: Step 1, the complex table data visualization system parses the table data uploaded by the user into a row-column data abstraction model; Step 2, the complex table data visualization system performs interactive transformation on the complex table data based on the row-column data abstraction model and table change operators; Step 3, the complex table data visualization system defines priorities for the table cells specified by the user and determines the visualization range cells based on the transformed complex table data and the row-column data abstraction model; Step 4, the complex table data visualization system converts the visualization range cells into a visualization form and embeds the visualization result into the complex table data. The method described in the present invention uses a flexible automatic recommendation mechanism for table visualization units, improving the efficiency of visualization construction.

Description

Technical Field

[0001] The present invention belongs to the technical field of visualization and human-computer interaction, and particularly relates to a method for data interactive transformation and visualization embedding of complex tables. Background Art

[0002] Tables are an important way of data management and are widely used in fields such as science, statistics, and finance. Visualizing raw table data helps users intuitively understand the data and improve data exploration efficiency. Existing table visualization technologies are mainly divided into two categories. The first category of technologies independently maps the data in a single cell to a visual element, and the second category integrates multiple adjacent cells together to construct a complex visualization form based on this data. Both of the above two types of table visualization technologies retain the structure and layout of the table to a certain extent, embed the visualization results in the table data, and enable users to avoid repeatedly switching between the original table data and the visualization results, thereby reducing their cognitive burden. However, existing research on table visualization only focuses on simple tables and has not considered complex tables with a hierarchical structure in the table header. Complex tables manage data more efficiently and are widely used in statistical reports and research papers, but existing table visualization technologies cannot support complex table data. First, the hierarchical structure of the table header makes there be a hierarchical relationship between the internal data of the table. For example, in complex table data, some cells represent the sum of data, and these cells are not comparable with non-sum data. In addition, some closely related data items may not be adjacent in a complex table, so it is impossible to construct a visualization form based on the scattered data items and embed it in the complex table.

[0003] For example, Chinese Patent Application No. CN201310001938.0 discloses an integrated method for automatic mapping using a database, including: Step 1, preparation for building a map database. Step 2, loading a vector map set and automatically building a database. Step 3, loading multi-frame images and multi-layer vector map data elements of the target area; integrated processing of images and vectors. Step 4, overlaying and displaying the loaded images and vector maps, drawing a data outsourcing bounding box and a map grid, and visually interactively selecting the map sheets to be cropped; Step 5, integrated batch cropping of images and vectors, batch cropping the loaded multi-frame images and multiple vector maps, and batch generating multiple map sheet cropping results; Step 6, integrated output of remote sensing images and vector map sets. This method can achieve integrated automatic mapping of remote sensing images and vector maps without frequent data format conversion and a large amount of repetitive labor. However, it does not propose a method for data interactive transformation and visualization embedding of complex tables. Figure 1 Integrated batch cropping of the body, batch cropping the loaded multi-frame images and multiple vector maps, and batch generating multiple map sheet cropping results; Step 6, integrated output of remote sensing images and vector map sets. This method does not require frequent data format conversion and can achieve integrated automatic mapping of remote sensing images and vector maps without a large amount of repetitive labor. Figure 1 However, it does not propose a method for data interactive transformation and visualization embedding of complex tables.

[0004] For another example, Chinese Patent Application No. CN201911071160.4 discloses a data intelligent analysis and visualization method based on big data, including: loading and importing all structured and unstructured data embedded with access connectors, as well as all data existing in various relational databases and various big data storage files into its built-in SPARK platform through a data connection module; this invention flexibly selects the modules and components required for different products according to the divided multiple independent functional modules to analyze the past, monitor the present and predict the future, and connect various different databases or distributed databases, supporting the connection and access to various big data storage platforms, merging, searching, visualizing and analyzing data, providing customers with one-stop storage and management services, helping customers calmly face the rapid growth of data and the risk of uncertain storage requirements of business systems, reducing customer risks while meeting the needs of business growth and changes. However, it also does not propose a method for interactive transformation and visualization embedding of complex tables. Summary of the Invention

[0005] To solve the above problems, the present invention provides a method for interactive transformation and visualization embedding of complex tables.

[0006] The method for interactive transformation and visualization embedding of complex tables includes:

[0007] Step 1: The complex table data visualization system parses the table data uploaded by the user into a row-column data abstraction model;

[0008] Step 2: The complex table data visualization system performs interactive transformation on the complex table data based on the row-column data abstraction model and table change operators;

[0009] Step 3: The complex table data visualization system defines priorities for the table cells specified by the user and determines the visualization range cells based on the transformed complex table data and the row-column data abstraction model;

[0010] Step 4: The complex table data visualization system converts the visualization range cells into a visualization form and embeds the visualization result into the complex table data.

[0011] Further, the complex table data visualization system in Step 1 parsing the table data uploaded by the user into a row-column data abstraction model includes:

[0012] Step 101: Based on the.xlsx data format and.csv data format uploaded by the user, the complex table data visualization system traverses the row list header table data with a hierarchical structure uploaded by the user, and parses and processes the complex table data according to the hierarchical structure of the table headers. When the name data of all the lower-level table headers corresponding to a certain level of table header are the same, the system abstracts the two-level table headers into a fully connected structure, and defines the table header with this characteristic as a double-cluster type structure. When the name data of all the lower-level table headers corresponding to a certain level of table header are different, it is defined as an independent hierarchical structure. In the same complex table, the double-cluster type structure and the independent hierarchical structure coexist;

[0013] Step 102: The complex table data visualization system constructs a row-column data abstraction model for the complex table data in the form of a combination of row list headers according to the double-cluster type structure and the independent hierarchical structure to determine the logical structure between the cells in the complex table data. The constructed row-column data abstraction model should have the ability to accurately describe the logical structure of the complex table and be applicable to complex table data of different structures, and have good expression ability for different logical structures; The complex table visualization system models all the table data according to the row list headers of the complex table. According to the characteristics of the two-dimensional table, the combination of the row and column headers is used as the unique identifier of the cell or cell block, that is, a cell is the intersection of a row and a column, and is represented by the combination of the row header and the column header; A cell block is a set of all cells within a rectangular range, and all the table data is represented by the combination of the corresponding multiple row headers and column headers.

[0014] Furthermore, the interactive transformation of the complex table data by the complex table data visualization system described in Step 2 based on the row-column data abstraction model and the table change operator includes:

[0015] Step 201: The complex table data visualization system identifies any operator of the user on the complex table data, including swap and transpose, linearization and stacking, fold and unfold operators;

[0016] Step 202: The complex table data visualization system adjusts the complex table data and the row-column data abstraction model according to the operator of the user on the complex table data.

[0017] Furthermore, the complex table data visualization system described in Step 3 defines the priority for the table cells specified by the user and determines the visualization range cells based on the transformed complex table data and the row-column data abstraction model, including:

[0018] Step 301: The complex table data visualization system identifies the area range specified by the user in the adjusted complex table data, and defines it as a table cell. The table cell includes two types, namely a single cell selected by the user and multiple consecutive cells selected by the user;

[0019] Step 302: The complex table data visualization system defines the calculation methods for the priorities of different table cell descriptors based on the adjusted row and column data abstraction model, including the priority calculation method based on the descriptor name and the priority calculation method based on the topological structure;

[0020] Step 303: The complex table data visualization system identifies the table cells selected by the user and calculates the priorities of the descriptors according to the priority calculation method selected by the user, calculating the row priority and the column priority respectively;

[0021] Step 304: When the complex table data visualization system identifies the table cells selected by the user and the priority determination method of the descriptors, it supports the user to separately select the recommended ranges of row and column priorities, and finally determines the visualization units.

[0022] Furthermore, the complex table data visualization system in step 4 converts the visualization range units into a visualization form and embeds the visualization result into the complex table data, including:

[0023] Step 401: The complex table data visualization system presets visualization templates based on the Vega-Lite descriptive grammar for the visualization range units, including visualization templates of unit visualization type, data overview type, trend tracking type, and correlation exploration type;

[0024] Step 402: The complex table data visualization system parses the visualization range units, specifies the visual mapping method according to the selected visualization template, and embeds it into the complex table data.

[0025] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0026] 1. In the complex table data interactive transformation and visualization embedding method of the complex table described in the present invention, the complex table visualization system constructs an abstract model by parsing the row and column headers of the complex table data. Based on the abstract model, the user interactively performs transformation operations on the complex table data. At the same time, the user selects a table cell in the complex table data, and the complex table visualization system recommends other relevant units related to the selected table cell by different priorities. According to the requirements, the priority range is selected and a series of target table cells are determined. Then, the user selects a visualization template and adjusts the configuration parameters, applies it to the target table cells, and continuously loops this process. Finally, the user completes the analysis and exploration process of the complex table data.

[0027] 2. The complex table data interactive transformation and visualization embedding method of the complex table described in the present invention directly uses the complex table with a hierarchical structure as the input data, automatically identifies various arbitrary-level tables, and establishes an abstract data model as the basis for data transformation and visualization recommendation.

[0028] 3. The data interactive transformation and visualization embedding method for complex tables according to the present invention supports rich data transformation operations, and directly interacts to change the relative positions of data and the length-width ratios of relevant data regions, facilitating users to directly embed rich and high-quality visualization forms in the original table.

[0029] 4. The data interactive transformation and visualization embedding method for complex tables according to the present invention provides flexible table cell declarations and a recommendation mechanism based on header names and topological structures based on an abstract model, supports users to select the application scope of visualization with a relatively high degree of freedom, enables users to avoid repeatedly selecting data regions and specifying visualization parameters, improves the exploration efficiency and reduces the operation burden.

[0030] 5. The data interactive transformation and visualization embedding method for complex tables according to the present invention supports constructing rich visualization forms with a single cell or a rectangular region composed of multiple adjacent cells as the basic unit, supports users to interactively explore and export the visualization results, and assists users in understanding the table data.

[0031] 6. The data interactive transformation and visualization embedding method for complex tables according to the present invention separates the logical structure and presentation form of complex tables by establishing a row-column data abstract model, and has two advantages: users directly and interactively operate the abstract model to change the presentation form of complex tables; the transformation operations of users on the tables are recorded in the abstract model, supporting the reuse of the abstract model. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 is a flowchart of the data interactive transformation and visualization embedding method for complex tables according to the present invention;

[0033] Figure 2 is a schematic diagram of the data of the complex table uploaded by the user and the complex table data visualization system parsing the table in the data interactive transformation and visualization embedding method for complex tables according to the present invention;

[0034] Figure 3 is a schematic diagram of the table data after three data transformation operations of exchanging, transposing, and linearizing the data of the original complex table by the user and the complex table visualization system parsing the table in the data interactive transformation and visualization embedding method for complex tables according to the present invention;

[0035] Figure 4 is a schematic diagram of two priority calculation method descriptions for describing the priorities of complex table data in the data interactive transformation and visualization embedding method for complex tables according to the present invention;

[0036] Figure 5Schematic diagram of the complex table data visualization system in the specific embodiment of the present invention for determining the visualization range unit based on the user's selection of table cells;

[0037] Figure 6 Analysis diagram of the complex table visualization system in the specific embodiment of the present invention for parsing the cells specified by the user in the complex table data. Specific embodiment

[0038] The present invention will be further described in detail below in conjunction with the specification drawings and specific embodiments.

[0039] As Figure 1 shown, the flowchart of the data interactive transformation and visualization embedding method for the complex table is as follows. The specific steps include:

[0040] Step 1: The complex table data visualization system parses the table data uploaded by the user into a row-column data abstraction model:

[0041] Step 101: According to the.xlsx data format and.csv data format uploaded by the user, the complex table data visualization system traverses the row-column header table data with a hierarchical structure uploaded by the user, and parses and processes the complex table data according to the hierarchical structure of the headers. When the name data of all the lower-level headers corresponding to a certain level of header are the same, the system abstracts the two-level headers into a fully connected structure, and defines the header with this characteristic as a double-cluster type structure. When the name data of all the lower-level headers corresponding to a certain level of header are different, it is defined as an independent hierarchical type structure. In the same complex table, the double-cluster type structure and the independent hierarchical type structure coexist;

[0042] Step 102: The complex table data visualization system constructs a row-column data abstraction model for the complex table data in the form of a combination of row-column headers according to the double-cluster type structure and the independent hierarchical structure to determine the logical structure between the cells in the complex table data. The constructed row-column data abstraction model should have the ability to accurately describe the logical structure of the complex table and be applicable to complex table data of different structures, and have good expression ability for different logical structures. The complex table visualization system models all the table data according to the row-column headers of the complex table. According to the characteristics of the two-dimensional table, the combination of the row and column headers is used as the unique identifier of the cell or cell block, that is, a cell is the intersection of a row and a column, and is represented by the combination of the row header and the column header; a cell block is a set of all the cells within a rectangular range, and all the table data is represented by the combination of the corresponding multiple row headers and column headers, as Figure 2 shown;

[0043] Step 2: The complex table data visualization system performs interactive transformation on the complex table data based on the row-column data abstraction model and the table change operator:

[0044] Step 201, the complex table data visualization system identifies any operator for complex table data, including swap and transpose, linearization and stacking, fold and unfold operators:

[0045] The user selects any operator to transform the complex table data. Among them, the six operators are divided into three groups according to their functions. The first group includes swap operation and transpose operation, which support changing the relative positions of headers at different levels and the corresponding data; the second group includes linearization operation and stacking operation, which do not change the relative positions of the existing headers in the original table, but add or delete specific headers and update the corresponding data items; the third group contains fold operation and unfold operation, which support the conversion between header content and table data;

[0046] Figure 3 Four data transformation operations for the original table data, namely swap, transpose, linearization, and fold operations, are given as follows Figure 3 -(a) represents the original complex table data, Figure 3 -(b)~ Figure 3 -(e) represent swap, transpose, linearization, and fold operations respectively, Figure 3 Stacking and unfold operations are not shown here. They are inverse operations of linearization 3-(d) and fold 3-(e) respectively;

[0047] The following is a detailed description of the six operators divided into three groups according to their functions:

[0048] (1) Swap and transpose:

[0049] The two transformation operations of swap and transpose only support table data with a double-cluster type structure header. Otherwise, these two transformations in complex table data will generate many meaningless null values, reducing the quality of the table data. The swap operation supports the user to change the order of two adjacent levels inside the row and column headers. The transpose operation supports the interchange between row and column headers. The user can select the appropriate header level for swapping according to different analysis requirements, so as to change the positions of relevant data, make comparable data adjacent in spatial positions, and support embedding visualization results in complex table data;

[0050] (2) Linearization and stacking:

[0051] Based on the constructed abstract model, the linearization and stacking operations operate on the data items in the table according to the header structure. The linearization operation calculates statistical values for the data items corresponding to all the lower-level headers of a certain layer of headers, including: sum, average, maximum, and minimum. The statistical values help users understand and analyze the overall trend of the data item set. However, these calculated data values are not comparable to other ordinary data, which hinders the effective analysis of users. For example, the sum is often much larger than ordinary data. If the two are visually encoded together without distinction, the visualization effect will be reduced. Here, a stacking operation that is inverse to the linearization is designed, allowing users to discard these derived headers and data values from the complex table;

[0052] (3) Folding and unfolding:

[0053] The folding and unfolding operations for complex table data support the mutual conversion of headers and the internal data of the table. The folding operation converts a row in the original complex table into multiple rows, folds multiple columns into one column, and copies the remaining columns, making the hierarchical table continuously flattened, allowing users to observe all the attribute values of a data item, such as Figure 3 shown in Figure 3 -(e - 1) and Figure 3 -(e - 2). Conversely, the unfolding operation builds a hierarchical structure on the basis of the flat data table, selects two columns of data and divides them into categorical or numerical types according to their characteristics, groups the numerical data by row according to the categorical data, and thus turns the categorical data into a new layer of headers. The unfolding operation supports distinguishing different attribute values in a column of data and rearranging these attributes and corresponding data in a rectangular space with a better aspect ratio, which is more suitable for creating various visualization forms;

[0054] Step 202, The complex table data visualization system adjusts the complex table data and the row and column data abstract model according to the operator of the user's operation on the complex table data;

[0055] Step 3, The complex table data visualization system defines priorities for the table cells specified by the user based on the transformed complex table data and the row and column data abstract model and determines the visualization range cells;

[0056] Step 301, The complex table data visualization system identifies the area range specified by the user in the adjusted complex table data and defines it as a table cell. The table cell includes two types, namely a single cell selected by the user and multiple consecutive cells selected by the user;

[0057] Step 302: The complex table data visualization system defines the calculation methods for the priorities of different table cell descriptors based on the adjusted row and column data abstraction model, including the priority calculation method based on the descriptor name and the priority calculation method based on the topological structure;

[0058] Figure 4 -(a) and Figure 4 -(b) show two priority calculation methods for the descriptors of table cells. Different numbers are used to represent the high and low priorities of table cell priorities, where the smaller the number, the higher the priority. The following introduces two specific calculation methods for descriptor priorities:

[0059] (1) Priority calculation method based on the descriptor name: Taking the row / column descriptor of the table cell selected by the user as the reference, the priority calculation method based on the descriptor name focuses on the name of the descriptor. All descriptors are divided into the following three categories: the reference descriptor selected by the user, the descriptors with the same name as the reference descriptor, and other descriptors. Among them, the reference descriptor has the highest priority of 0, the descriptors with the same name have a priority of 1, and other descriptors have a priority of 2.

[0060] (2) Priority calculation method based on the topological structure: Taking the row / column descriptor of the table cell selected by the user as the reference, the priority calculation method based on the topological structure focuses on the topological structure information of the descriptor in the row and column abstract data model. Find the level of the nearest common ancestor descriptor of all descriptors and the reference descriptor selected by the user, and subtract the level number of the reference descriptor from the level number of the nearest common ancestor to obtain the priority of the required descriptor.

[0061] After calculating the priorities of the row and column descriptors of the table cells respectively using the above two specific calculation methods, the row and column priorities of all table cells are obtained. Figure 4 Different colors in it correspond to different priorities;

[0062] Step 303: The complex table data visualization system identifies the table cell selected by the user and calculates the priorities of the descriptors according to the priority calculation method selected by the user, calculating the row priority and the column priority respectively;

[0063] Step 304: When the complex table data visualization system identifies the table cell selected by the user and the priority determination method of the descriptor, it supports the user to separately select the recommended range of row and column priorities, and finally determines the visualization unit. Based on the priorities defined by colors in Figure 4 it gives the table cells recommended according to the priority range specified by the user. Figure 5 Figure 5- (a) shows different colors when the user selects the data within the virtual frame and when the user selects the priority range: the row range is [0, 1], and the column range is [0, 1, 2]. At the same time, the visualization system recommends the cell range that can be visualized. Similarly, Figure 5 - (b) shows the case where the row range is [0, 1, 2] and the column range is [0, 1];

[0064] Step 4: The complex table data visualization system converts the visualization range cells into a visualization form and embeds the visualization result into the complex table data:

[0065] Step 401: The complex table data visualization system presets visualization templates based on the Vega-Lite descriptive grammar for the visualization range cells, including visualization templates for cell visualization type, data overview type, trend tracking type, and correlation exploration type. Regarding the cell visualization type in the visualization template of Step 401, it only applies to table cells composed of a single cell, while the other three types of visualization templates for data overview type, trend tracking type, and correlation exploration type apply to table cells containing multiple cells:

[0066] (1) The cell visualization type visualizes the value of each cell and shows the relative size between comparable cells without changing the original table structure. Users can use different visual channels (such as color, size, and position, etc.) to encode the value of each cell to facilitate the perception of the relative size of the table data. For example, by choosing color as the visual encoding, users can construct a visualization result similar to a heatmap to understand the data;

[0067] (2) The data overview type is used to display the values of all cells in the table cells. According to whether the values in the table cells are aggregated, the data overview type visualization is divided into two categories: non-aggregated type and aggregated type. The non-aggregated type technology encodes each value in the table cell into a visual element and calculates the position according to the value of the cell, while the aggregated type calculates a single statistical value based on multiple values in the row / column and then encodes it into the visual element;

[0068] (3) The trend tracking type supports users to understand the change of values when the column or row changes. The trend tracking type encodes the values of a row or a column and its corresponding header sequence to support users to understand the trend of data change. The corresponding visualization forms mainly include horizontal line charts and line charts;

[0069] (4) Correlation exploration type allows users to explore the correlation between multiple rows or columns in a table cell, and the corresponding visualization forms mainly include scatter plots and heat maps. Different from the above-mentioned other visualization forms such as cell visualization type, data overview type, and trend tracking, this type of visualization form does not bind categorical data to the coordinate axes, but binds numerical data to the coordinate axes. Specifically, users can encode the data of different rows or columns to different coordinate axes respectively to verify whether there is a correlation between them;

[0070] Step 402, the complex table data visualization system analyzes the visualization range unit, specifies the visual mapping method according to the selected visualization template and embeds it into the complex table data. Figure 6 The decomposition of the table cell is given, and the decomposition result of the table cell contains three parts: x-categorical data, y-categorical data, and numerical data.

[0071] It should be noted that after the user selects the visualization template preset by the complex table data visualization system, it is necessary to specify the visual mapping method of the data based on the selected visualization template. To enhance the intuitiveness of the user's construction of the visualization result, the complex table data visualization system provides visual mapping configuration options. Most visualization forms contain one or more axes arranged horizontally or vertically. To ensure the positional correspondence between the visual elements and the original data, the mapping rules designed by the complex table data visualization system only allow users to bind x-categorical data to the horizontal axis or y-categorical data to the vertical axis. In addition, the complex table data visualization system designs feasible parameters for visual mapping, realizing the reuse of visualization parameter settings. For the visualization of complex tables, users need to apply the same visualization parameter configuration to all target table cells according to the abstract model mentioned in step 1 and the recommended results mentioned in step 3. The results obtained after each table cell is deconstructed according to step 3 are different, and the visualization parameter configuration should also be adjusted accordingly. The visualization mapping mechanism of the complex table data visualization system decouples the visualization configuration from its corresponding underlying data and uses the method of calculating the offset to determine the visualization parameter configuration of different table cells.

[0072] To evaluate the effectiveness of the present invention for analyzing complex table data with a hierarchical structure, an existing visualization software tool Tableau similar to the function of the present application was selected as a reference benchmark to support the transformation of table data and the construction of visualization forms.

[0073] In the experiment, artificially generated tabular data was used to better control the progress of the experiment. To help users more easily understand the data and the analysis tasks being performed, an experimental dataset was generated with the background of time - city industrial production GDP. In this table data, the row headers are "City" and "Industrial Category", while the column headers are "Year" and "Quarter". Based on this framework, multiple datasets were generated. These datasets have the same size and structure, but the specific values of the data are different.

[0074] This application selected two tasks for comparative research, including Task 1 to obtain an overview of the data and Task 2 to explore data details. Among them, Task 1 requires users to understand the changing trend of GDP over time for each city. Users need to analyze and identify the cities where the GDP change trend is different from other cities. While Task 2 requires users to analyze the total GDP of different cities and the GDP values of each industry. First, users need to determine the year when the total city GDP reaches the maximum value. Then, compare the proportion of GDP of each industry in that year. Finally, analyze and obtain the industry with the largest proportion.

[0075] By recruiting people who often use tables to organize and analyze data, they used this application and the visualization software tool Tableau to complete the above two tasks, and recorded their results and completion times respectively. At the end of the experiment, by analyzing the time and accuracy of users completing each task, according to the results of the t - test, the process of using this application for both tasks was significantly faster than the visualization software tool Tableau. As for accuracy, the accuracy rate of using the visualization software tool Tableau to complete Task 1 was 78.57%, and the accuracy rate of using this application to complete it was 92.86%. The accuracy rate of using the visualization software tool Tableau to complete Task 2 was 64.29%, and the accuracy rate of using this application to complete it was 85.71%.

[0076] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the scope disclosed by the present invention should be covered by the protection scope of the claims of the present invention.

Claims

1. A method for interactive transformation and visualization embedding of complex table data, characterized in that: include: Step 1: The complex table data visualization system parses the table data uploaded by the user into a row and column data abstract model; Step 1 also includes: Step 101, according to the .xlsx data format and .csv data format uploaded by the user, the complex table data visualization system traverses the row and column header table data with a hierarchical structure uploaded by the user, and parses and processes the complex table data according to the hierarchical structure of the header. When the name data of all lower-level headers corresponding to a certain layer of headers are the same, the system abstracts the two layers of headers into a fully connected structure, and defines the header with this feature as a dual-cluster type structure. When the name data of all lower-level headers corresponding to a certain layer of headers are different, it is defined as an independent hierarchical structure. In the same complex table, the dual-cluster type structure and the independent hierarchical structure exist at the same time; Step 102, the complex table data visualization system constructs a row and column data abstract model for the complex table data in the form of a combination of row and column headers according to the dual-type structure and the independent hierarchical structure, so as to determine the logical structure between cells in the complex table data. The constructed row and column data abstract model should have the ability to accurately describe the logical structure of the complex table and be applicable to complex table data of different structures, and have good expression capabilities for different logical structures; the complex table visualization system models all table data according to the row and column headers of the complex table. According to the characteristics of the two-dimensional table, the combination of row and column headers is used as the unique identifier of the cell or cell block, that is, a cell is the intersection of a row and a column, and is represented by a combination of row headers and column headers; a cell block is a collection of all cells within a rectangular range, and all table data are represented by a combination of corresponding multiple row headers and column headers; Step 2: The complex table data visualization system interactively transforms the complex table data based on the row and column data abstract model and the table change operator; Step 3: The complex table data visualization system defines the priority of the table cells specified by the user and determines the visualization range cells based on the transformed complex table data and the row and column data abstract model; Step 4: The complex table data visualization system converts the visualization range cells into a visualization form and embeds the visualization results into the complex table data.

2. The interactive transformation and visualization embedding method of complex table data according to claim 1, characterized in that: In step 2, the complex table data visualization system interactively transforms the complex table data based on the row and column data abstract model and the table change operator, including: Step 201, the complex table data visualization system identifies any operator of the user on the complex table data, including exchange and transposition, linearization and stacking, folding and unfolding operators; Step 202: The complex table data visualization system adjusts the complex table data and row and column data abstract models according to the user's operators on the complex table data.

3. The interactive transformation and visualization embedding method of complex table data according to claim 1, characterized in that: In step 3, the complex table data visualization system defines the priority of the table cells specified by the user and determines the visualization range cells based on the transformed complex table data and the row and column data abstract model, including: Step 301: The complex table data visualization system identifies the area range specified by the user in the adjusted complex table data and defines it as a table cell. The table cell includes two types, namely, a single cell selected by the user and multiple continuous cells selected by the user. Step 302: The complex table data visualization system defines different priority calculation methods for table cell descriptors based on the adjusted row and column data abstract model, including a priority calculation method based on a descriptor name and a priority calculation method based on a topological structure. Step 303, the complex table data visualization system identifies the table cell selected by the user, and calculates the priority of the descriptor according to the priority calculation method selected by the user, and calculates the row priority and the column priority respectively; Step 304: When the complex table data visualization system identifies the table unit and descriptor priority calculation method selected by the user, it supports the user to select the recommended row and column priority ranges respectively, and finally determines the visualization unit.

4. The interactive transformation and visualization embedding method of complex table data according to claim 1, characterized in that: In step 4, the complex table data visualization system converts the visualization range unit into a visualization form and embeds the visualization result into the complex table data, including: Step 401, the complex table data visualization system presets visualization templates based on Vega-Lite descriptive syntax for visualization range units, including visualization templates of unit visualization type, data overview type, trend tracking type, and correlation exploration type; Step 402: The complex table data visualization system parses the visualization range unit, specifies the visual mapping method according to the visualization template selected by the user, and embeds it into the complex table data.

Images