Data Insight in DataWorks DataAnalysis provides deep insights into your data, supporting advanced data exploration and visualization. You can use Data Insight to understand data distributions, create data cards, and assemble them into a data report. The results can also be shared as a single-image report. Powered by artificial intelligence (AI), this feature helps you parse complex data and make informed business decisions.
Prerequisites
Permissions: You must have the required permissions for data analysis. For a list of supported workspace roles, see Preset role permissions for DataAnalysis.
For information about how to grant permissions to a role, see Add a workspace member and manage member role permissions.
Region limitations: This feature is available only in the following regions: China (Hangzhou), China (Shanghai), China (Shenzhen), China (Chengdu), China (Zhangjiakou), China (Beijing), and China (Ulanqab).
Accessing Data Insight
Log on to DataWorks DataAnalysis, switch to the target region, and click Go to DataAnalysis. In the left-side navigation pane, click the 
icon to open the Data Insight page.
Create a data insight task
You can create a data insight task in multiple ways: from the Data Insight list, from a table in the SQL Query data catalog, or from the results of an SQL query.
From the Data Insight list
On the Data Insight page, click Create Data Insight Task. In the Create Data Insight Task dialog box, configure the table object for insight, and click Start Data Insight.
The data source and data range default to your last used settings, with the most recent partition selected automatically.
The configuration items include workspace, data source type, data source name, access identity mode (executor identity or default access identity of the data source), table name, field, and data range (Random 10,000 rows or full dataset). You can also select Use large model for computation.
(Optional) If you have not enabled enhanced analysis, click Try Now as prompted.
From a table in SQL Query
This feature is available only in the legacy version of DataAnalysis and is not supported in the new version.
In the left-side navigation pane, click the
icon to open the SQL Query page and add the table to the catalog on the left.In the top toolbar, click the catalog list icon to open the Add Catalog panel. Select the Recommended Catalog tab. Next to the target table, such as My MaxCompute Tables, click Add to Catalog.
In the data catalog on the left, find the table to analyze.
On the table basic information page, click Data Insight.
Right-click the table name and select Data Insight from the shortcut menu.
In the Create Data Insight Task dialog box, the parameters are automatically populated. Click Start Data Insight.
(Optional) If you have not enabled enhanced analysis, click Try Now as prompted.
From SQL query results
New version
Hover over , and then click
> New SQL File.Write your SQL code. In the toolbar at the top of the SQL editor, click Run and wait for the results.
In the results, click the
icon, select a Data Range, and then click Start Data Insight.(Optional) If you have not enabled enhanced analysis, click Try Now as prompted.
Legacy version
In the left-side navigation pane, click the
icon to open the SQL Query page. In My Files, click Create File, run your SQL query, and wait for the results.In the results, click the
icon, select a Data Range, and then click Start Data Insight.In the data range, select Only data displayed in the query result area, and select the fields to analyze, such as rank, cate_id, cate_name, total_sales_amount, order_count, and pt.
(Optional) If you have not enabled enhanced analysis, click Try Now as prompted.
Data insight results
Wait for the data insight task to complete.
NoteYou can click Stop Analysis at any time to terminate the task.
Running data exploration on a full dataset may take a long time. We recommend that you check the results later on the Data Insight list page. Closing the current page does not interrupt the ongoing exploration process.
The Data Insight page shows the progress of a three-step workflow: Data Preparation (automatic data ingestion), Data Exploration (statistical distribution, high-frequency data ratio, max/min/average values, etc.), and Generate Cards (saving exploration conclusions and recommending data cards). These steps run sequentially. A Stop Running button is available in the upper-right corner.
View data insight results: The system automatically generates visualization chart cards based on data types and other characteristics.
After cards are generated, click load more to display additional cards.
At the top of the Data Insight results page, a collapsible Hide statistics and filters panel displays cards showing the number of unique values and data distribution for each field. For example, in the behavior_type field,
pvaccounts for 98% andbuyaccounts for 2%. If there are many fields, you can use the pagination arrows (< and >) to the right of the card area to browse more field statistics cards. Below the panel, a data table shows the raw detailed data.If the data range is not set to the full dataset, hover over the upper-right corner of a chart to save it to Cards as needed.
FAQ
Q: What causes the insight result to show "No more cards"?
A: The dataset is too small to generate insight results. For best results, use a dataset with more than 500 rows.
Q: How do I troubleshoot a failed table-based data insight task?
A: Ensure that you can connect to and successfully query the data source. You can verify this by running a query in SQL Query first.