Skip to content
DashboardGPT

What the semantic model is and how to generate one

It tells the assistant which columns are metrics and dimensions.

For the complete AI-readable content index, see /llms.txt.

Semantic model · 1 min read

A semantic model tells DashboardGPT which columns are metrics (numbers to aggregate, such as revenue) and which are dimensions (things to group by, such as region or month). Better models mean better answers.

  1. 1Open Semantic model in the sidebar (or the Semantic model tab on a dataset).
  2. 2Choose a dataset that is Ready or Published.
  3. 3If it has no model yet, choose “Generate semantic model”. DashboardGPT builds one from the column profile.
  4. 4Review the metrics and dimensions, then choose “Publish”.

Related reading

Guides from Learn on the why and when behind this.

Still stuck?

Ask the docs assistant about “What the semantic model is and how to generate one” or anything else in DashboardGPT.