CN-121979947-A - Interactive multidimensional data visual display sand table and method for business data analysis

CN121979947ACN 121979947 ACN121979947 ACN 121979947ACN-121979947-A

Abstract

The application provides an interactive multi-dimensional data visual display sand table and a method for business data analysis, which are used for carrying out semantic mapping on state parameters in interactive control operation to obtain a plurality of visual intention metadata and further generating a structured query instruction when receiving the interactive control operation, extracting a target data subset of heterogeneous business data source visual rendering from a multi-dimensional data cube of a business data source based on the structured query instruction, carrying out visual optimization on a chart type, a visual coding channel and initial visual parameters of visual rendering through a base number of a dimension field, a data distribution type of a measurement field and a numerical range in the target data subset to obtain a visual configuration set of the target data subset, and using the visual configuration set to render to generate a multi-dimensional data visual view and outputting and displaying the data visual view in a graphical user interface. Based on the scheme, the closed loop can be automatically generated from heterogeneous data to interactive views.

Inventors

SONG AIYUN
LI PENGJING
LI JUNWEN
LUO JIE

Assignees

桂林理工大学南宁分校

Dates

Publication Date: 20260505
Application Date: 20251224

Claims (10)

1. An interactive multi-dimensional data visualization method for business data analysis, comprising the steps of: Acquiring an original data record in a heterogeneous commercial data source, identifying and standardizing dimension fields and measurement fields in the original data record, and carrying out multi-layer aggregation on all dimension fields through each measurement field to construct a multidimensional data cube of the heterogeneous commercial data source in a memory; When receiving an interactive control operation, carrying out semantic mapping on state parameters in the interactive control operation to obtain a plurality of visual intention metadata, and further converting all the visual intention metadata into a structured query instruction containing target dimensions, target metrics, expected chart types and data screening conditions; Extracting a target data subset of heterogeneous commercial data source visual rendering from the multidimensional data cube based on the structured query instruction, and performing visual optimization on the chart type, the visual coding channel and the initial visual parameters of visual rendering through the base numbers of dimension fields, the data distribution types and the numerical ranges of measurement fields in the target data subset to obtain a visual configuration set of the target data subset; and rendering and generating a multi-dimensional data visual view by using the visual configuration set, and outputting and displaying the data visual view in a graphical user interface.
2. The method of claim 1, wherein identifying and normalizing the dimension fields and the metric fields in the raw data record specifically comprises: Reading structured metadata and unstructured sample data of an original data record, and identifying candidate dimension fields and measurement fields based on a preset field naming rule base and a data pattern inference algorithm; carrying out semantic disambiguation and category merging on the candidate dimension fields, unifying field definitions of synonymous but different names, and establishing a dimension hierarchical relationship according to the domain knowledge graph to obtain dimension fields; And carrying out data type verification and unit standardization conversion on the candidate measurement fields, unifying all numerical measurement to a reference measurement unit, and marking the aggregation function types of the measurement units to obtain the measurement fields.
3. The method of claim 1, wherein the multi-layer aggregation of all dimension fields by each metric field, and further wherein constructing a multi-dimensional data cube of a heterogeneous business data source in memory comprises: According to the dimension hierarchy relation, precalculating an aggregation index table of a plurality of granularity levels for each dimension field, wherein the aggregation index table supports the operations of rolling up and down along the hierarchy; Based on the aggregate function type of the measurement field, calculating and caching corresponding measurement summary values and statistical characteristic values for each dimension combination granularity in a memory in advance; and establishing a star logic model between the dimension and the measurement through each aggregation index table, all measurement summary values and statistical characteristic values to form a multidimensional data cube supporting online processing operation.
4. The method of claim 1, wherein semantically mapping the state parameters in the interactive control operation to obtain the plurality of visual intention metadata specifically comprises: capturing all state parameters in the interaction control operation; Mapping each state parameter to a corresponding dimension field, a metric field and a screening conditional expression in the multidimensional data cube to form an original intention parameter which can be understood by a machine; Deducing the potential analysis intention of the user according to the combination logic and the business rule of the control, and supplementing and generating implicit analysis dimension and analysis measurement; and carrying out intention fusion on the original intention parameters, the implicit analysis dimension and the analysis measurement to form a plurality of visual intention metadata.
5. The method of claim 1, wherein converting all visual intention metadata into structured query instructions comprising target dimensions, target metrics, desired chart types, and data filtering conditions specifically comprises: performing conflict detection and priority arbitration on each visual intention metadata, and when the intention of mutual exclusion or redundancy exists, performing trade-off or combination according to a preset rule; The intention parameters in the visual intention metadata after arbitration are converted into standard query statement fragments, and the target dimension, target measurement and data screening conditions of the query are defined; based on the user history preferences and the visualized global settings, a plurality of recommended expected chart types are specified for the query results, and then the target dimensions, the target metrics, the expected chart types and the data screening conditions are packaged into a structured query instruction.
6. The method of claim 1, wherein extracting a target data subset of a heterogeneous business data source visualization rendering from the multidimensional data cube based on the structured query instruction comprises: applying the screening conditions in the structured query instruction to the corresponding dimensions of the multidimensional data cube, and determining the data range slice of the query; positioning corresponding pre-aggregate data blocks in the multidimensional data cube according to the target dimension and the target measurement in the structured query instruction, and triggering real-time aggregation calculation; and slicing the data range, aggregating the calculated results in real time, organizing a result set according to the format and sequence required by the query, and attaching context metadata to form a target data subset for visual rendering.
7. The method of claim 1, wherein visually optimizing the chart type, the visual coding channel and the initial visual parameters of the visual rendering through the cardinality of the dimension field, the data distribution type of the metric field and the numerical range in the target data subset, to obtain the visual configuration set of the target data subset specifically comprises: automatically determining the roles of the target data subset in the chart as facets, coloring or coordinate axes according to the base numbers of the dimension fields in the target data subset; matching graph types showing distribution characteristics by using the data distribution types and the numerical ranges of each measurement field in the target data subset; and allocating a proper visual coding channel for the selected chart type according to a visual perception effectiveness principle, optimizing initial visual parameters based on the numerical range of each measurement field, and generating a visual configuration set of the target data subset.
8. An interactive multidimensional data visualization presentation sand table for business data analysis, comprising: The acquisition module is used for acquiring an original data record in the heterogeneous commercial data source, further identifying and standardizing dimension fields and measurement fields in the original data record, carrying out multi-layer aggregation on all dimension fields through each measurement field, and further constructing a multidimensional data cube of the heterogeneous commercial data source in a memory; The processing module is used for carrying out semantic mapping on state parameters in the interaction control operation when the interaction control operation is received, so as to obtain a plurality of visual intention metadata, and further converting all the visual intention metadata into a structured query instruction containing target dimensions, target metrics, expected chart types and data screening conditions; The processing module is further used for extracting a target data subset of the heterogeneous commercial data source visual rendering from the multidimensional data cube based on the structured query instruction, and performing visual optimization on the chart type, the visual coding channel and the initial visual parameters of the visual rendering through the base numbers of the dimension fields and the data distribution types and the numerical ranges of the measurement fields in the target data subset to obtain a visual configuration set of the target data subset; and the execution module is used for generating a multi-dimensional data visual view by using the visual configuration set rendering and outputting and displaying the data visual view in a graphical user interface.
9. A computer device comprising a memory for storing a computer program and a processor for calling and running the computer program from the memory, such that the computer device performs the interactive multi-dimensional data visualization method for business data analysis of any one of claims 1 to 7.
10. A computer readable storage medium having instructions or code stored therein which, when executed on a computer, cause the computer to implement the interactive multi-dimensional data visualization method for business data analysis of any one of claims 1 to 7.

Description

Interactive multidimensional data visual display sand table and method for business data analysis Technical Field The application relates to the technical field of data visualization, in particular to an interactive multi-dimensional data visualization display sand table and method for business data analysis. Background The data visualization display sand table is an enterprise analysis platform integrating multi-source data processing, a real-time analysis engine and interactive graphic rendering. The method supports the user to dynamically explore and simulate the scene of multidimensional data through natural interaction such as dragging, screening, drilling and the like. The method is characterized in that complex data are converted into visual objects which can be intuitively operated, an immersive analysis environment which is free in layout and multi-view linkage and can record analysis paths is formed, so that professional analysis threshold is reduced, and decision insight efficiency and cooperation capability are improved. In a traditional technical architecture, building a business data analysis system first faces a high degree of manual reliance on model definition and data pipelines. The data engineer needs to manually design a star/snowflake data model based on the prior understanding of the service scene, explicitly specify a dimension table, a fact table structure and an association relation, and meanwhile, must write a complex script to process format differences, value range unification and quality cleaning of heterogeneous data sources, which is time-consuming and further causes the service logic to be solidified in the code and table structure too early, when the service requirement changes, the prior art scheme often needs to trace back and modify ETL logic, reconstruct a data model, re-run historical data and even adjust a downstream report semantic layer. The chain type dependence ensures that the system has long response changing period and high cost, is difficult to adapt to the business analysis scene of agile iteration, and forms the fundamental contradiction between the evolution of the business demand and the realization rigidity of the technology. Therefore, how to automatically generate a closed loop from heterogeneous data to interactive views, thereby improving the intuitiveness of business data analysis has become a difficult problem for the industry. Disclosure of Invention The application provides an interactive multi-dimensional data visual display sand table and a method for business data analysis, which can realize automatic generation of a closed loop from heterogeneous data to interactive views, thereby improving the intuitiveness of business data analysis. In a first aspect, the present application provides an interactive multi-dimensional data visualization method for business data analysis, comprising: Acquiring an original data record in a heterogeneous commercial data source, identifying and standardizing dimension fields and measurement fields in the original data record, and carrying out multi-layer aggregation on all dimension fields through each measurement field to construct a multidimensional data cube of the heterogeneous commercial data source in a memory; When receiving an interactive control operation, carrying out semantic mapping on state parameters in the interactive control operation to obtain a plurality of visual intention metadata, and further converting all the visual intention metadata into a structured query instruction containing target dimensions, target metrics, expected chart types and data screening conditions; Extracting a target data subset of heterogeneous commercial data source visual rendering from the multidimensional data cube based on the structured query instruction, and performing visual optimization on the chart type, the visual coding channel and the initial visual parameters of visual rendering through the base numbers of dimension fields, the data distribution types and the numerical ranges of measurement fields in the target data subset to obtain a visual configuration set of the target data subset; and rendering and generating a multi-dimensional data visual view by using the visual configuration set, and outputting and displaying the data visual view in a graphical user interface. In some embodiments, identifying and normalizing the dimension and metric fields in the raw data record specifically includes: Reading structured metadata and unstructured sample data of an original data record, and identifying candidate dimension fields and measurement fields based on a preset field naming rule base and a data pattern inference algorithm; carrying out semantic disambiguation and category merging on the candidate dimension fields, unifying field definitions of synonymous but different names, and establishing a dimension hierarchical relationship according to the domain knowledge graph to obtain dimension fields; And carrying ou