Google Takeout + Spotify API feed a BigQuery warehouse. Python handles extraction/loading + enrichment calls, while dbt handles transformations + tests to produce analytics marts consumed by Power BI. Documentation and delivery are tracked in Notion + Jira.

Source: infographic adapted and reworked from Le Wagon – https://www.lewagon.com/fr

Core data flow (1 glance)

Data Sources

→ Extract & Load (Python)

→ BigQuery — Raw

→ Transform (dbt: staging → intermediate → marts + tests)

→ BI (Power BI)

→ (later) Orchestrate (n8n + dbt Cloud jobs) + Docs (Notion & Jira)

Data sources

1) Google Takeout

• music-library.csv → library dataset (Sprint 1)
• watch-history.json → listening history extraction (Sprint 2)

Used to enrich tracks with metadata such as genres/sub-genres, with reusable matching logic across Library + History.

Responsibilities: Python vs dbt (clear boundary)