The Three Layers
The medallion architecture organises data into progressively cleaner and more business-ready layers:
- Bronze (Raw): Raw data ingested directly from source systems with no transformation. Append-only. Serves as the single source of truth and enables reprocessing.
- Silver (Cleansed): Deduplicated, validated, and lightly transformed data. Conformed data types, null handling, and basic business rules applied.
- Gold (Curated): Aggregated, business-domain-specific datasets ready for BI tools and data science. Wide, denormalised tables optimised for query performance.