Business intelligence (often referred to as BI or analytics) has evolved a great deal in this release; many elements of BI that were devolved to external applications are now first class citizens of the development environment. We will now define dimensions and measurements in an Operations project that are built and released as any other Operations project. Our role in this is to provide the basis which the business intelligence designer will use to develop powerful business intelligence solutions.

When developing a BI solution, it is usual to create a new project that references the packages it analyzes. This allows the solutions to be developed at the same time, and also allows the analytics and other projects to be deployed independently. Of course, if the data structures change in a way that breaks the analytics solution, the build server will usually highlight this.

When planning this type of work, the analytics should be designed along with the main solution design and not as an afterthought, even though we will start analytics later in the project lifecycle. This results in the data structures being correctly designed with analytics and reporting in mind; when we come to develop the analytics element of the project, we should be able to do so without changes to the underlying tables. Changes to underlying tables could be considered like adding a basement garage to your house.

The recipes in this chapter are based on creating an analytics project for customer invoice analysis.