Friday, April 20, 2012

Big Data and Dimensional Modeling [Video]

I am often asked, "Is the dimensional model obsolete in the world of big data?"

I always respond with a question of my own: "Do you plan to stop measuring your business?"  Of course, the answer is always "no." 

Dimensional modeling is all about measurement, and so it will continue to have a place in our BI programs.

Here, I discuss the topic with Paul Kautza, Director of Education at The Data Warehousing Institute (TDWI).

In this interview, I talk about four ways we use dimensional models:
  1. To capture business requirements (e.g. measurement definitions)
  2. To define and manage a roadmap for a BI program (scope)
  3. As the basis for a database design (star schema, snowflake schema, cube)
  4. As a presentation layer for the business (semantic layer)
Developments in the world of big data may impact the third category, but not always.