Building the Unified Star Schema

The Data (or Information) Mart used in a Data Vault implementation often use the Kimball Dimensional Modelling techniques. The star schema approach is easy to use. But the Kimball approach has a big limitation: it does address what happens when a query involves multiple fact tables. How do we merge two star schemas? It usually results in an “ad-hoc approach”, where a data analyst writes an SQL query, specific for a business requirement. But this means that the business users do not have full access to a self-service analytics experience: for every new requirement they need to go through a data expert. What if we could build a special data structure that involves all the required facts and dimensions, all connected together in a single star schema, easy to use, and always available for self-service analytics?

Subsequent to this presentation Francesco and Bill’s book “The Unified Star Schema: An Agile and Resilient Approach to Data Warehouse and Analytics Design” is now available on Amazon and at https://technicspub.com/uss/.