The Team
The team in an agile project is more effective if it is small and cross-functional. Five to ten people is a good size.
The project team for an agile BI project should consist of people with all the necessary skills and domain knowledge to complete the project. This includes:
- Understanding of the problem that is being solved
- Understanding of the source applications and data sources
- Understanding of user-facing BI tools and technologies
- Understanding of BI architectures
- Understanding of data transformation and storage
In order to satisfy these needs the team might be comprised of:
- End user: To represent the eventual consumers someone who will be an end-user of the system is better than a proxy.
- Business analyst: This person understands
- IT developer: This person understands the organization's IT infrastructure, platforms, and standards.
- Database administrator: This person knows about data extraction, data manipulation, and databases. This person also understands the data structures of the source applications.
- BI specialist: This person has experience of the BI toolset being used and knows how to install, configure, and use the tools.
Location
If at all possible the project team members should sit together. The ideal environment is a room with plenty of white-board and flip-chart room and a large table for the team to sit at. If it is not possible to get the team out of cube-land having them sit in adjacent cubes will greatly help communication. Since the team members are cross-functional it is likely that some or all of them will have to move from their usual locations.
End-To-End Participation
The iterations of an agile project might have a shifting set of tasks. Since some of these tasks are functionally specific some of the team members will different levels of involvement as the project moves from phase to phase.
In an agile project the team as a whole commits to creating and delivering the features of each iteration.
Meetings
Deliverables
Acceptance
Top-Down, Bottom-Up and Meet-In-The-Middle BI
Top-Down BI Process
- Design cubes, business views
- Auto generate / manually create star schema
- ETL / populate star schema
Bottom-Up BI Process
- Design star schema or start with an already existing database
- ETL / populate star schema
- auto generate / manually create cubes and business views
Meet-in-the-middle BI Process = Refactoring