Customer Analytics Pipeline
Generated from notebooks/ by Claude via the Datadef MCP server.
Architecture

What the notebooks do
| Notebook | Reads | Writes |
|---|---|---|
01_ingest_events.ipynb | S3 raw events | bronze.events |
02_sessionise.ipynb | bronze.events | silver.sessions |
03_customer_features.ipynb | silver.sessions | gold.customer_features |
Notes
- Sessionisation uses a 30-minute inactivity window.
gold.customer_featuresis the only table the ML jobs read.- The diagram above is embedded from the canvas, so it updates when the pipeline does — no re-export step.









