The Evolution of Data Architecture
summary
Claims that a data warehouse is a persistent, subject‑oriented store separating transactional and analytical worlds, created by heavy ETL work that normalizes things like gender codes into a single corporate representation. Shows a real spiderweb systems diagram representing ~10,000 people and explains why dozens of proliferating data marts produce conflicting answers. Counters quick Kimball‑style marts with the slow, integration‑heavy path to data Vault and a cloud‑hosted future that converges structured and unstructured into a data lakehouse.
AI-written summary · about
chapters (11)
- 0:00Data architecture evolution
- 1:53Spiderweb systems and disintegrity
- 4:03Data warehouse solution
- 6:52ETL and transformation challenges
- 8:40Data warehouse to marts
- 10:08Structured data limitations
- 14:47Need for unstructured data
- 17:40Q&A on warehouse models
- 22:04Hardest warehouse implementation
- 24:11Organizational pain and architect role
- 27:01Conclusion
related talks
42mBuilding Tomorrow's Warehouse Today: Modern Data Warehouse Modelling Insights
31mData Warehouses Meet Data Lakes
31mData Unleashed: A Developer's Perspective on Navigating the Architecture Maze
58mTales of Data Architecture Evolution
01h 01mModern Data Warehousing Strategy
01h 00mBuilding an Effective Data Warehouse Architecture
05mThe Future of Data Warehousing
41mModern Data Architectures – What They Are and How to Implement?