Roadmap › ETL & Data Warehousing Concepts › Module: Dimensional Modelling
Lessons in this course (12)
8. Inmon vs Kimball, data marts, data lakes and lakehouses
IntermediateMediumLesson 8 of 12
- Inmon (top-down) — build a normalised enterprise data warehouse (3NF) first, then dependent data marts.
- Kimball (bottom-up) — build dimensional data marts per business process with conformed dimensions (the "bus matrix") — together they form the warehouse.
- Data mart — subject-specific subset (Sales mart, Finance mart).
- Data lake — cheap storage of raw data in any format (files: Parquet, CSV, JSON) — schema-on-read.
- Lakehouse — data lake + warehouse features (ACID tables via Delta/Iceberg, SQL, governance) — e.g. Databricks, Microsoft Fabric OneLake.
- Data Vault 2.0 — hubs (business keys), links (relationships), satellites (attributes with history); agile, auditable raw warehouse layer.