Bronze, silver, gold: the medallion architecture
Raw data in, trusted data out, in three layers.
The idea
Most lakehouses organize data in three layers. Bronze keeps raw data exactly as it arrived. Silver cleans, deduplicates and types it. Gold shapes it for a business question, like daily revenue by region.
The layers make problems traceable: when a dashboard number looks wrong, you can walk back through gold and silver to the raw record that caused it.
What the lesson will build
- The same orders data stored as bronze, silver and gold Parquet files
- The transformation between each layer, and the tests that guard it
- A gold table answering one business question
Key ideas
- Raw versus cleaned versus modeled data
- Lineage
- Where to put data quality checks
- Lakehouse table formats
The video
- Long form: Three stacked layers filling one after another, with a single bad record caught between bronze and silver.
- Short: Bronze, silver, gold. Each with one job.
When it’s published, the code will live in data-architecture/ and this page will link to it.