Chapter 3. Creating the data mart

This chapter is recommended for

Mondrian makes it easy for users to do analysis, but behind the scenes it requires data organized in a way that’s convenient for analysis. Historically data has been organized for operational use in third normal form (3NF), but Mondrian has adopted the use of star schema structures based on industry best practices.

In this chapter, we’ll cover the general architecture of an analytic solution and then explore star schemas, the “best practice” database modeling technique for analytic systems. We’ll dig into their specifics, understanding that Mondrian is expecting to perform its analytic magic on top of a star schema. We’ll compare this with third normal form modeling and examine some of the high-level benefits of the star schema for an analytic system.

We’ll conclude with a few additional aspects of the star schema technique, including how to manage changes to data over time, modeling the all-important time dimensions.

By the end of this chapter, you’ll understand how data is structured to make analysis with Mondrian possible. This chapter is primarily aimed at the data architect, but other readers will likely find understanding the data architecture useful as well.

3.1. Structuring data for analytics

3.2. Additional star schema modeling techniques

3.3. Summary