Automatically drawing infographics for statistical data based on a data model

Patent metadata

Field	Value
Publication number	US-12299007-B2
Application number	US-202217581646-A
Country	US
Kind code	B2
Filing date	Jan 21, 2022
Priority date	Jan 21, 2022
Publication date	May 13, 2025
Grant date	May 13, 2025

Abstract

A computer-implemented method, system and computer program product for automatically drawing infographics. Variables of a dataset are received from a computing device that were selected by the user of the computing device. For those selected variables that are associated with a data model, a procedure to draw infographics for variables assigned or not assigned the role of a target using the data model associated with each of the variables assigned or not assigned the role of target, respectively, is implemented. Alternatively, if the selected variables are not associated with a data model, then such variables are assigned a level of measurement as well as assigned the role of input. Such assignments become the data model which, along with the metadata (e.g., values of the variable) obtained by parsing the original data, are used to implement the procedure to draw infographics for variables not assigned the role of a target.

First claim

The invention claimed is:

1. A computer-implemented method for automatically drawing infographics, the method comprising:
populating a database with variables and associated data models based on previous assignments of levels of measurements and roles of variables;
receiving selected variables from a dataset to be utilized for generating infographics from a computing device by an infographics generator;
utilizing said data models stored in said database connected to said infographics generator for classifications of variables used in datasets, wherein said classifications of variables used in datasets comprise a type of variables, a role of variables, and a level of measurement of variables;
identifying an appropriate data model stored in said database based on matching a name of a variable selected by a user with a name of a variable listed in said data model using natural language processing;
parsing, by said infographics generator, said dataset to obtain metadata;
automatically implementing a procedure to draw infographics, by said infographics generator, for variables not assigned a role of a target using said metadata and said data model associated with each of said variables not assigned said role of said target in response to said variables not being assigned said role of said target;
automatically implementing a procedure to draw infographics, by said infographics generator, for variables assigned said role of said target using said metadata and said data model associated with each of said variables assigned said role of said target in response to said variables being assigned said role of said target;
determining if there is a value of a continuous variable that exceeds a first threshold value; and
performing a pair-group search strategy to reduce a complexity of correlation analysis and in selecting appropriate continuous variables to be drawn thereby improving clarity and understandability of infographics in response to said value of said continuous variable exceeding said first threshold value, wherein said pair-group search strategy comprises:
grouping continuous variables with a value that exceeds said first threshold value;
computing a correlation rate for each pair of continuous variables by computing a Euclidean distance or a cosine distance between said pair of continuous variables, wherein said correlation rate measures show strong a relationship is between two variables;
identifying a pair of continuous variables as belonging to a same cluster in response to said computed correlation rate exceeding said first threshold value; and
identifying a pair of continuous variables as not belonging to said same cluster in response to said computed correlation rate not exceeding said first threshold value.

2. The method as recited in claim 1 further comprising:
determining whether a value for each of said received variables not associated with said data model is less than a second threshold value;
assigning a level of measurement to be continuous for a first variable not associated with said data model without having a value that is less than said second threshold value;
assigning a level of measurement to be nominal for a second variable not associated with said data model having a value that is less than said second threshold value; and
assigning each of said received variables not associated with said data model to have a role of input.

3. The method as recited in claim 1 , wherein said procedure to draw infographics for variables not assigned said role of said target comprises one or more of the following selected from the group consisting of: drawing a relationship map for variables with assigned levels of measurement of nominal and ordinal, drawing a histogram for each variables with an assigned level of measurement of continuous, drawing a pie graph for each variable with an assigned level of measurement of nominal, and drawing a bar graph for each variable with an assigned level of measurement of ordinal.

4. The method as recited in claim 1 , wherein said procedure to draw infographics for variables not assigned said role of said target comprises identifying clusters of variables to be used to draw a scatter plot in response to said value of continuous variables exceeding said first threshold value.

5. The method as recited in claim 1 , comprising:
wherein said data model includes classifications of variables used in datasets, wherein said classifications comprise a type of variable, a role of said variable, and the level of measurement of said variable.

6. The method as recited in claim 1 further comprising:
selecting a variable in said same cluster to be paired in a new group with a variable not belonging to said same cluster in response to not comparing all continuous variables belonging to said same cluster or not belonging to said same cluster.

7. The method as recited in claim 1 further comprising:
forming a data model for variables not previously associated with a data model using a level of measurement and a role assigned to said variables.

8. A computer program product for automatically drawing infographics, the computer program product comprising one or more computer readable storage mediums having program code embodied therewith, the program code comprising programming instructions for:
populating a database with variables and associated data models based on previous assignments of levels of measurements and roles of variables;
receiving selected variables from a dataset to be utilized for generating infographics from a computing device by an infographics generator;
utilizing said data models stored in said database connected to said infographics generator for classifications of variables used in datasets, wherein said classifications of variables used in datasets comprise a type of variables, a role of variables, and a level of measurement of variables;
identifying an appropriate data model stored in said database based on matching a name of a variable selected by a user with a name of a variable listed in said data model using natural language processing;
parsing, by said infographics generator, said dataset to obtain metadata;
automatically implementing a procedure to draw infographics, by said infographics generator, for variables not assigned a role of a target using said metadata and said data model associated with each of said variables not assigned said role of said target in response to said variables not being assigned said role of said target;
automatically implementing a procedure to draw infographics, by said infographics generator, for variables assigned said role of said target using said metadata and said data model associated with each of said variables assigned said role of said target in response to said variables being assigned said role of said target;
determining if there is a value of a continuous variable that exceeds a first threshold value; and
performing a pair-group search strategy to reduce a complexity of correlation analysis and in selecting appropriate continuous variables to be drawn thereby improving clarity and understandability of infographics in response to said value of said continuous variable exceeding said first threshold value, wherein said pair-group search strategy comprises:
grouping continuous variables with a value that exceeds said first threshold value;
computing a correlation rate for each pair of continuous variables by computing a Euclidean distance or a cosine distance between said pair of continuous variables, wherein said correlation rate measures show strong a relationship is between two variables;
identifying a pair of continuous variables as belonging to a same cluster in response to said computed correlation rate exceeding said first th

Assignees

• IBM

Inventors

• Fan Ye
• Mao Qi
• Wu Juan
• Wu Jia Zhong
• Fan Long
• Liu Chong
• Yu wen pei
• Yang Yang

Classifications

• G06T11/26 Primary

  Drawing of charts or graphs · CPC title

• G06F16/287 Primary

  Visualization; Browsing · CPC title

• G06T11/206

  Physics · mapped topic