The data visualization feature in DataWind is called visual query, which is an important step from data to decision-making.
Users can perform self-service visual analysis based on the connected data without needing to know SQL syntax. By clicking or dragging, users can generate visual charts and gain data insights. This module supports a variety of chart and analysis settings, flexible filtering based on dimensions, dates, metrics, and dynamic dimensions/dynamic metrics, interactive analysis such as drill-up and drill-down, chart linkage, as well as intelligent analysis operations like attribution analysis, empowering data-driven decision-making and business growth.

As shown in the figure above, the visual query editing interface provides users with a streamlined set of operations. The interface sections are introduced as follows:

1. Field list area: Displays the list of fields in the dataset used by the current visual query. The field list includes dataset fields and personal fields. Users can switch datasets in this area and more.

2. Chart and analysis configuration area:

   • Chart configuration area: Provides configuration options for each chart, such as color settings, X-axis settings, Y-axis settings, and more.
   • Analysis configuration area: Provides analysis functions such as sorting, Top N, total, percentage, comparison, year-over-year, month-over-month, reference lines, and table calculations.

3. Field configuration area: Field types include dimensions and metrics. Dragging a field into this area allows further configuration of a specific field to complete the query.

4. Filter configuration area: Provides filtering capabilities for dimension and metric fields.

5. Chart display area: Displays the chart style according to the various query conditions configured by the user, providing a WYSIWYG experience.

6. Overall configuration area (top toolbar): In this top toolbar area, you can undo or redo operations, and options are provided for viewing query history, downloading, monitoring and alerts, as well as query configuration. The following describes the specific operations for the overall configuration area:

   Names of operation items in the overall configuration area	Description
   Undo	Return to the previous operation and cancel the most recent configuration change.
   Redo	Restore the undone operation, for restoring after an accidental undo.
   Reset	Restore the current query or configuration to its initial state and clear all temporary changes.
   Query history	Click the Query history button to view the historical query records of the current dataset, making it easy to trace or reuse previous queries.
   Download	Click the Download button, select Download Data, Download Image, or Configure Download as needed from the drop-down list, configure the selected option accordingly, and click the Download button to download the relevant content.
   Monitoring	Click the monitoring chart to select Create Monitoring or open the monitoring list. warning After setting the pivot table - tree display to incremental fetching, monitoring is not supported; existing monitoring cannot be pushed.
   Query configuration	Click the Query Configuration button to configure the query. Select whether to enable caching. If enabled, select the cache validity period. Select cluster query parameters from the drop-down list. After completing the configuration, click the Apply button.
   Rename	Modify the name of the current chart.
   Chart list	Click the Chart List button to go to Chart Management in the Project Center and view the chart list.
   Save as	Save the current configuration as a new copy.
   Save	Save the current configuration or changes to the system, overwriting the previous version.

Set dimension and metric conditions #

For example, select province and year-month as dimensions, and select GMV as the metric (you can drag fields from the field list on the left to the data panel to select them, or click the field to have it appear in the dimension or metric row of the data panel).

Select chart type #

To display richer charts, such as a bar chart, click the bar chart icon on the chart panel to select a bar chart.

Set sorting #

To sort, click sorting on the analysis panel.

Save as chart #

Click the Save button on the toolbar to save as a chart. If a dashboard has already been created, you can directly select the saved dashboard to quickly add the chart to the dashboard; if no dashboard has been created or you do not intend to display it on a dashboard, you can leave it unselected. You can later find this chart in Chart Management.

Field #

In the field list area, you can view and use the fields of the current dataset. Visual query provides five field types: string, integer, float, date, and datetime. For detailed usage and more, see Overview of Field Configuration.

Dimensions and metrics #

A field has two roles: dimension and metric. A dimension is used to segment the data being analyzed and determines the perspective from which to view the data. It is a non-aggregated field, such as date or author UID. A metric measures data using different aggregation methods and is usually a number, such as the number of submissions by an author. For example, to view the number of students under each course category, set "Course Category" as the dimension and "Number of Students" (that is, the sum of students) as the metric.

Application of dimensions and metrics in different business scenarios #

A field may serve as a dimension or a metric in different scenarios. Detail table demo:

• Each time an author submits once, one record is generated.
• Filtered 2,230 rows of data from 2020-1-6 to 2020-1-8.

Scenario 1 Number of submissions per day View the total number of submissions for each day by date

• Dimension: Date
• Metric: Number of submissions, aggregation method: sum

Scenario 2 Number of submissions per author per day View the total number of submissions by different authors for each day by date and author

• Dimension: Date, Author UID
• Metric: Number of submissions, aggregation method: sum

Scenario 3 Number of submitting authors Number of authors who submitted between 2020-1-6 and 2020-1-8

• Metric: Author UID, aggregation method: distinct count

Aggregation of metrics #

The product provides multiple aggregation methods. For details, see 2.1 Set aggregation method.
The following shows the values obtained using different aggregation methods for a sample detail table (containing 7 rows of data).

Results returned by different aggregation methods:

Query and auto query #

• For extracted datasets, auto query is enabled by default
• For direct connection datasets, auto query is disabled by default
  For the differences between extraction and direct connection, see: Connection methods.

Auto query can be enabled at the top of the chart display area. If auto query is enabled, each time you change the query conditions or chart configuration, the visualization chart will refresh; if auto query is not enabled, the chart will be grayed out after each change, and you need to click the "Query" button to re-initiate the query. When the amount of query data is large, to avoid wasting resources and waiting for queries after each minor change, it is recommended to disable auto query.