BigQuery

What you need

Item	Description
Project ID	Your Google Cloud project ID (e.g., `my-company-prod`)
Dataset ID	The BigQuery dataset containing your event tables (e.g., `analytics`)
Service account credentials	A JSON key file for a service account with `BigQuery Data Viewer` and `BigQuery Job User` roles

Create the service account

If you do not already have a service account with the right permissions:

Go to the Google Cloud Console IAM page.
Create a new service account (e.g., levered-reader).
Grant it the BigQuery Data Viewer and BigQuery Job User roles on the project.
Create a JSON key and download it.

Connect via the dashboard

Navigate to Settings > Warehouse in the Levered dashboard.
Select BigQuery as the provider.
Enter your Project ID and Dataset ID.
Upload your service account credentials JSON file.
Click Test Connection to verify Levered can reach your warehouse.
Click Save.

Troubleshooting

Connection test fails

Confirm the service account has BigQuery Data Viewer and BigQuery Job User roles. Make sure the project ID and dataset ID are correct and the dataset exists.

Permission errors at training time

Levered runs SQL queries during model training. If training fails with permission errors, check that the service account has SELECT access on the specific tables referenced in your metric queries.

Next step

Now that your warehouse is connected, set up your data pipeline -- create the tables Levered reads during training.