Data quality improvement actions

The quality of data directly affects the quality of decision making. Good quality data provides business process insights, a better understanding of customers, and faster problem identification. Data quality actions identify problems that you should address to improve data quality in your data estate.

Required roles

• To see data quality actions in your environment, users need the data quality reader role.
• To assign and modify actions, users need the data quality steward role.

Data quality action types

The types of data quality actions generated are:

• A rule that falls below default thresholds.
• A global data quality score that falls below the threshold.
• A data quality scanning and data quality profiling job that fails.
• A data quality scanning and data quality profiling job that skips due to no changes on data since the last run.
• Outlier presence in profiled columns.
• Too many nulls in the profiled data asset.

Data quality actions severity

Data quality actions have a severity level and color indicator depending on the data quality score:

• High (red): Data quality score is less than 25.
• Medium (orange): Data quality score is less than 50.
• Low (yellow): Data quality score is less than 75.

View and manage data quality actions for governance domains

Data quality actions roll up from the data asset level to the governance domain level. To browse actions and assign them to users for further action:

1. In Microsoft Purview Unified Catalog, select Health Management, then select Data quality.

2. Select a governance domain from the list.

3. The Action items button shows the number of action items for the domain. Select Action items to view all data quality actions.

4. Select an action from the list of actions to view the action's details page.

5. From here, you can assign the action to a user to investigate and resolve. The Recommendation section suggests the action to take and provides queries to find the low quality data.

   Note

   • Not every action has a SQL query because not all actions are data related or related to a data quality rule. Some actions are system related, such as a failed or skipped data quality job.
   • If a custom rule exception occurs, which means that the data quality score generated from the custom rule doesn't meet the defined threshold, an action is created. However, a SQL query isn't suggested.

   

6. Update the Status*, then select Save.

View and manage data quality actions for data assets

You can browse data quality actions for a data asset and assign them to users from the data quality overview page:

1. In Microsoft Purview Unified Catalog, select the Health Management menu and Data quality submenu.

2. Select a governance domain from the list.

3. Select a data product from the list, then select a data asset from the list of assets.

4. On the data asset's details page, select View all on the Actions tile to see all actions for the data asset.

5. Browse through the list of Active, In progress, Resolved, and My items actions.

   

6. Select an action to see its details and assign it to a user to investigate and resolve. an individual for remedy. The Recommendation section suggests the action to take and provides queries to find the low quality data.

   Note

   • Not every action has a SQL query because not all actions are data related or related to a data quality rule. Some actions are system related, such as a failed or skipped data quality job.
   • If a custom rule exception occurs, which means the data quality score generated from the custom rule doesn't meet the defined threshold, an action is created. However, a SQL query isn't suggested.

   

7. Update the Status, then select Save.

Resources

• Best practices for data quality.