Notebooks¶

Interactive data exploration with Marimo notebooks and DuckDB.

Overview¶

Dashboards show you what's happening — notebooks let you ask why. Dango integrates Marimo, a reactive Python notebook, with your DuckDB warehouse so you can run ad-hoc queries, build visualizations, and investigate data issues without leaving the platform.

Key features:

• Reactive Python notebooks — Marimo cells re-execute automatically when dependencies change
• DuckDB snapshots — notebooks query a read-only copy of your warehouse, so syncs and transforms are never blocked (see DuckDB core concepts)
• File locking — time-limited locks with heartbeat prevent conflicts when multiple users edit notebooks
• Starter templates — pre-built notebooks for data exploration and quality checks

How It Works¶

When you open a notebook, Dango performs these steps automatically:

1. Lock — acquires a file-level lock so no one else can edit the same notebook
2. Snapshot — copies data/warehouse.duckdb to .dango/snapshots/ (if Marimo isn't already running)
3. Start — launches the Marimo server (headless, on port 7805)
4. Connect — opens the notebook in your browser, connected to the read-only snapshot
5. Release — when you're done (Ctrl+C or release via web UI), the lock is released

Quick Start¶

# Create a notebook from the explore template
dango notebook new --template explore --name my_analysis

# Open it in Marimo
dango notebook open my_analysis

# Press Ctrl+C to release the lock and exit

Key Concepts¶

CLI vs Web UI¶

# List notebooks
dango notebook

# Create a notebook
dango notebook new --template explore --name my_analysis

# Open a notebook (locks, snapshots, starts Marimo)
dango notebook open my_analysis

Next Steps¶