Establishing 'Making it Happen' as a 'Formal & Predictable' Discipline

	Data Mart Dimension Fact table Matrix in Data Warehouse Dimensional Model

ENCYCLOPEDIA→ Enterprise Intelligence → - Data-Warehouse/Mart → - DW Dimensional Modeling Process →

Data Mart Business Theme Matrix in Data Warehouse Dimensional Model

The start point of business themes and the end delivery of data marts have to be linked to confirm that we have not missed any business themes.

Before we proceed, lets understand on what do we mean by Data-Mart here. Data-Mart here is one single Data Warehouse schema [a combination of interlinked fact table(s) and dimension table(s)] in the overall Data Warehouse design. Each Data Warehouse can have many such schemas. Please refer dimensional model and Schemas of a dimensional model to get a better picture. These Data-Marts (OR schemas) could be then residing in any layer of Data Warehouse OR Data-Mart OR both, depending upon the Data Warehouse Architecture Scenarios you will use.

Pick-up all the business themes, requirements and specifications documents, and create a matrix placing the business themes on one axis and the data marts on the other. As you go through this page, you may open up Execution-MiHPractice Tool Business Themes+ Data Mart matrix on the side to understand how this looks like.

The matrix will contain:

All the Data Warehouse business themes (within and out of phase I and II).
All the Data Marts titles.
The Descriptive objectives of the Data Marts.
Phase of the Data Mart.

Making a list of Data Marts

There is no single fixed way of defining the set of Data Marts. One of the ways to list a Data Mart can be taken as a three cornered view.

View 1 – With the list of Measures, make sets of measures (facts). A Mart should be a centered around a set of Measures. For example:

No. of sales units
Value of sales units
Cost of sales Units
This Data Mart can be called a Sales Analysis Data Mart.

View 2 – Even with the given set of Facts, a business theme can drive a separate set of Data Marts. For example, the same set of above matrix could drive two different data marts. Each of these Data Marts will have the above set of facts and also the facts specific to that Mart. For example-

The Sales Data Mart having additional facts – Commission Value, Average Shelf time
The Production Data Mart having additional Data facts – Order to Delivery time, Productivity.

View 3 – Understanding of the source Data.

In the initial stages, it is advised to create data marts having minimum number of data sources. You may like to design each data-mart around a limited set of systems.

TIP- Typically a Data Mart can have multiple business themes, but ideally you should not have a business theme split across multiple data marts. This is because typically you will have one data-mart equivalent to one cube (refer OLAP). Your BI tool may not allow you to run a query across two cubes at the same time. Even if it does, the configuration is little complex.

REFER Execution-MiHPractice Tool Business Themes+ Data Mart matrix