# US 2019/0377728 A1 · Method and system for data analysis with visualization

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2019/0377728 A1
- Document ID: US20190377728A1
- Status: Published application
- Kind: A1
- Application number: 16/246,906
- Filed: 2019-01-14
- Published: 2019-12-12
- Earliest claimed: 2018-06-06
- Assignees: Shanghai Development Center of Computer Software Technology
- Applicants: Shanghai Development Center of Computer Software Technology
- Inventors: Lizhi Cai; Mingang Chen; Wenjie Chen; Zhenyu Liu; Yun Hu; Jianhua Wu; Wei Song; Dali Chen; Binliang Wu; Lianghe Ling
- Web page: https://patentgrep.com/patent/US20190377728A1

## CPC

- G06F 16/248: Presentation of query results
- G06F 16/26: Visual data mining; Browsing structured data
- G06F 16/24539: {using cached or materialised query results}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A method and system are provided for data analysis with visualization. The method includes; generating a visual result; generating a second visual result according to the second query condition and the visual parameter; generating a recommended query condition; and generating a final visual result according to the recommended query condition selected by the user. By using the analysis method or system in the present invention, a new query in which the user may be interested is generated based on an original user query, to guide the user to quickly understand knowledge hidden in the data. An analysis result is presented to the user visually, and is more visual, clearer, and easier to understand compared with a numerical calculation result. In addition, the result can be displayed by using a variety of graphics.

## Claims

### Claim 1

A data analysis method with visualization, wherein the analysis method comprises:
obtaining to-be-analyzed data;
obtaining a data format and a first query condition that are defined by a user;
generating a first visual result according to the data format and the first query condition that are defined by the user and the to-be-analyzed data;
obtaining a second query condition and a visual parameter that are defined by the user, wherein the visual parameter comprises a visual type, a visual data display range, a visual color, and a visual size;
generating a second visual result according to the second query condition and the visual parameter that are defined by the user and the first visual result;
generating a recommended query condition according to a historical query condition by using a recommendation algorithm, for the user to perform selection, wherein the historical query condition is a query condition used prior to the second query condition, and the historical query condition comprises the first query condition; and
generating a final visual result according to the recommended query condition selected by the user and the second visual result.

### Claim 2 (depends on 1)

The analysis method according to claim 1, wherein the generating a first visual result according to the data format and the first query condition that are defined by the user and the to-be-analyzed data specifically comprises:
performing field segmentation on the to-be-analyzed data according to the data format, to obtain segmented data;
correcting the segmented data to obtain corrected data;
filtering data, corresponding to the first query condition, in the corrected data according to the first query condition, to obtain filtered data; and
generating the first visual result based on the filtered data.

### Claim 3 (depends on 2)

The analysis method according to claim 2, wherein the generating a second visual result according to the second query condition and the visual parameter that are defined by the user and the first visual result specifically comprises:
filtering data, corresponding to the second query condition, in the corrected data according to the second query condition, to obtain twice-filtered data; and
generating the second visual result according to the twice-filtered data and the visual parameter.

### Claim 4 (depends on 1)

The analysis method according to claim 1, wherein the first visual result comprises a histogram, a pie chart, a broken line chart, an area graph, a scatter diagram, a bar chart, a bubble diagram, a curve fitting chart, a box plot, a jean chart, a matrix graph, a map, a parallel coordinate chart, a radar map, a word cloud chart, and a user-defined visual effect chart.

### Claim 5 (depends on 1)

The analysis method according to claim 1, wherein after the generating a second visual result, the method further comprises:
storing the first query condition to a set of the historical query condition.

### Claim 6 (depends on 1)

The analysis method according to claim 1, wherein the generating a recommended query condition according to a historical query condition by using a recommendation algorithm specifically comprises:
obtaining a correlation matrix R between all attributes of the to-be-analyzed data according to a Pearson correlation coefficient algorithm, wherein:
R
=
[
1
r
12
…
r
1

n
r
21
1
…
r
2

n
…
…
…
…
r
n


1
r
n


2
…
1
]
;
a set of all the attributes of the to-be-analyzed data is (α1,α2, . . . , rij is a Pearson correlation coefficient between an attribute αi and an attribute αj, i=1,2, . . . , and j=1,2, . . . ,
calculating, according to a formula σj=min rij, a recommendation level σj of an attribute αj that does not exist in a historical query, wherein αi is an attribute that has existed in the historical query;
successively obtaining recommendation levels of all attributes that do not exist in the historical query, to obtain a recommendation level set;
sorting elements in the recommendation level set by value, to obtain an element with a smallest value;
determining an attribute that is corresponding to the element with a smallest value and that does not exist in the historical query, as a recommended attribute; and
adding the recommended attribute to the second query condition and generating the recommended query condition.

### Claim 7

A data analysis system with visualization, wherein the analysis system comprises:
a to-be-analyzed data obtaining module, configured to obtain to-be-analyzed data;
a user-defined data obtaining module, configured to obtain a data format and a first query condition that are defined by a user;
a first visual result generation module configured to generate a first visual result according to the data format and the first query condition that are defined by the user and the to-be-analyzed data;
a user interaction module, configured to obtain a second query condition and a visual parameter that are defined by the user, wherein the visual parameter comprises a visual type, a visual data display range, a visual color, and a visual size;
a second visual result generation module configured to generate a second visual result according to the second query condition and the visual parameter that are defined by the user and the first visual result;
a recommended-query-condition generation module, configured to generate a recommended query condition according to a historical query condition by using a recommendation algorithm, for the user to perform selection, wherein the historical query condition is a query condition used prior to the second query condition, and the historical query condition comprises the first query condition; and
a final visual result generation module configured to generate a final visual result according to the recommended query condition selected by the user and the second visual result.

### Claim 8 (depends on 7)

The analysis system according to claim 7, wherein the first visual result generation module specifically comprises:
a segmentation unit, configured to perform field segmentation on the to-be-analyzed data according to the data format, to obtain segmented data;
a correction unit, configured to correct the segmented data, to obtain corrected data;
a filtering unit, configured to filter data, corresponding to the first query condition, in the corrected data according to the first query condition, to obtain filtered data; and
a first visual result generation unit configured to generate a first visual result based on the filtered data.

### Claim 9 (depends on 8)

The analysis system according to claim 8, wherein the second visual result generation module specifically comprises:
a second filtering unit, configured to filter data, corresponding to the second query condition, in the corrected data according to the second query condition, to obtain twice-filtered data; and
a second visual result generation unit configured to generate the second visual result according to the twice-filtered data and the visual parameter.

### Claim 10 (depends on 7)

The analysis system according to claim 7, wherein the recommended query condition generation module specifically comprises:
a correlation matrix obtaining unit, configured to obtain a correlation matrix R between all attributes of the to-be-analyzed data according to a Pearson correlation coefficient algorithm, wherein:
R
=
[
1
r
12
…
r
1

n
r
21
1
…
r
2

n
…
…
…
…
r
n


1
r
n


2
…
1
]
;
a set of all the attributes of the to-be-analyzed data is (α1,α2, . . . , rij is a Pearson correlation coefficient between an attribute αi and an attribute αj, i=1,2, . . . , and j=1,2, . . . ;
a recommendation level calculation unit, configured to calculate, according to a formula σj=min rij, a recommendation level σj of an attribute αi that does not exist in a historical query, wherein αi is an attribute that has existed in the historical query;
a recommendation level set obtaining unit, configured to successively obtain recommendation levels of all attributes that do not exist in the historical query, to obtain a recommendation level set;
a sorting unit, configured to sort elements in the recommendation level set by value, to obtain an element with a smallest value;
a recommended attribute determining unit, configured to determine an attribute that is corresponding to the element with a smallest value and that does not exist in the historical query, as a recommended attribute; and
a recommended query condition generation unit, configured to add the recommended attribute to the second query condition, to generate the recommended query condition.