Medallion Architecture Vs. Star Schema! Where Data Modeling Fits

Medallion architecture and star schema answer different questions, so you never have to pick one. Medallion says how refined data is. A star schema is the shape the gold layer usually takes. We follow one order, SO10482 (two mugs and a tin of Earl Grey, $37.99), from a raw JSON payload in bronze to two fact table rows in gold. Silver deduplicates it, casts the types and conforms one customer from two systems into a single ID. Gold declares the grain and builds the star with surrogate keys and a degenerate dimension. Then we cover what else gold can hold (one big table, pre-aggregated tables, ML feature tables), how the layers line up with dbt's staging, intermediate and marts and with Kimball's and Inmon's architectures, and the four mistakes teams make when they let the layers decide the model. ⏱️ Chapters 0:00 - Medallion Vs. star schema: not rivals 0:58 - Medallion architecture in one minute 2:12 - What a star schema is: facts, dimensions, grain 3:28 - Why the two get confused 4:49 - One order through bronze, silver and gold 5:19 - Bronze: land it as is 5:44 - Silver: dedupe, cast, conform 6:31 - Gold: declare the grain, build the star 7:24 - What else gold can hold 9:05 - Mapping the layers to dbt, Kimball and Inmon 10:42 - 4 mistakes when the layers decide the model 12:07 - The verdict 🔗 Links Databricks, What is the medallion lakehouse architecture?: https://docs.databricks.com/aws/en/lakehouse/medallion Microsoft Fabric: https://www.microsoft.com/microsoft-fabric dbt: https://www.getdbt.com #medallionarchitecture #starschema #medallionvsstarschema #dimensionalmodeling #datamodeling #Databricks #lakehouse #bronzesilvergold #facttable #Kimball #dbt #MicrosoftFabric #dataengineering #dataengineeringforbeginners