Dataset materialization for AI/BI dashboards for faster loading

Dataset materialization pre-computes and stores data beforehand, allowing published dashboards to render from saved results instead of rerunning expensive queries on every interaction.

Use cases

Imagine you have a dashboard with multiple filters and complex joins that takes too long to load on every user click.

• Speed up heavy dashboards without remodeling data.
• Reduce warehouse costs on repeated reads.
• Automate creation and storage without managing pipelines or schemas.

How to use

To enable materialization, configure the option when publishing your dashboard:

1. Open the dashboard publication side menu.
2. Select the shared data permissions option.
3. Turn on the dataset materialization toggle.
4. Click publish.

Limitations and requirements

• Feature is in Beta.
• Requires workspace admins to enable it from the previews page.
• Works only on published dashboards using shared data permissions.
• Datasets with parameters are not supported.
• Data protected by row-level security or column masks cannot use materialization.

Using it in a pipeline

Apply materialization at the end of the dashboard delivery flow to optimize consumption in the presentation layer (Gold), avoiding constant reprocessing of large tables.