Skip to Content
Data StudioNotebooks

Notebooks

Notebooks are an interactive transformation type for exploring and analyzing marketing data with Python. Vendo supports both browser-side execution (via Pyodide) for instant feedback and server-side runs for production workloads.

Create a notebook from Data → Transformations → New Data Model → Notebook.

Browser-side execution (Pyodide)

For exploration and prototyping, notebooks run Python directly in your browser:

  • pandas and numpy available out of the box
  • A built-in vendo helper: vendo.read_bq(sql) queries your warehouse and returns a DataFrame, and vendo.show(data) renders tables, charts, and values as cell output
  • No server round-trip for cell execution — instant feedback
  • Ideal for ad-hoc analysis, data inspection, and prototyping transformations

Browser-side runs do not write to production tables. Use them to develop logic before scheduling a server-side run.

Note: The notebook environment is distinct from the saved Python transformation contract. Notebooks let you use the vendo API and free-form cells; saved Python models run a flat, no-import script with injected input tables that assigns its result to output.

Server-side execution

For production workloads, schedule a server-side notebook run:

  • Executes against your warehouse with full compute resources
  • Writes outputs to configured tables or output targets
  • Supports scheduling (daily, weekly, on sync completion)

Workflow

  1. Create a Notebook transformation.
  2. Write cells with Python — explore data with pandas.
  3. Test locally in the browser (Pyodide).
  4. When ready, configure server-side output and schedule.
  5. Monitor runs from the transformation detail page.
Last updated on
Grouping RulesSegments