Data Warehouse Concepts
Central repository for analytical data.
Architecture
Source systems → ETL → Staging → Warehouse → Data Marts. Kimball vs Inmon methodologies.
Star schema: fact table + dimension tables. Snowflake: normalized dimensions.
Technologies
Snowflake, BigQuery, Redshift: cloud data warehouses. Columnar storage for analytics. Massive parallelism.
ETL vs ELT
ETL: transform before load. ELT: load raw, transform in warehouse. Modern prefers ELT.
Key Takeaways
- Data warehouse centralizes analytical data
- Star schema common for analytics
- Cloud warehouses provide scalability