Dividing the World

DM Review, March, 2008 by Ralph Kimball

In my November, December and January columns and the data quality article published in the October issue of DM Review, I laid a solid foundation for building a data warehouse. We have done a careful job of gathering all the overlapping design constraints; we have established a good set of boundaries with all the groups we interact with; we have captured a perfect subset of changed data to feed our data extraction; and we have built a powerful architecture for cleansing the data once it is under our control.

Our next big task is to divide the data into dimensions and facts. We call these designs dimensional models. Dimensions are the basic stable entities in our environment, such as customers, products, locations, marketing promotions and calendars. Facts are the numeric...