My first Publication Agile-Data-Warehouse-Design-eBook | Page 123

102
Chapter 4
Scan down the
matrix columns to
identify potential
conformed
dimensions
Scan across the
matrix rows to
compare process
complexity
Start with a high-
level matrix to help
you plan
dimensionally
Scanning down the dimension columns reveals the potential for dimensional
conformance. Conformed dimensions that could form a data warehouse bus show
up with multiple ticks. The contrast between these valuable dimensions that
support cross-process analysis (Hurray!) and the non-conformed dimensions that
do not (Boo!) should encourage everyone to work towards conformance.
Scanning across the process rows helps to estimate the complexity of a business
process: generally, the more dimension ticks, the more complex a process is likely
to be and the more resource needed to define its business events and implement
them.
It’s a good idea to start your agile DW/BI project by creating a high-level matrix to
help you plan your data warehouse design from a conformed dimensional process
measurement perspective from the outset. You may want to add to it some of the
additional features of the event matrix described below.
Use a high-level dimensional matrix to gain support from senior business and IT
management for conforming dimensions.
The Event Matrix
The event matrix is
a modelstorming
version of the
dimensional matrix
It contains details
for BEAM✲ event
An event matrix is a more detailed version of the dimensional matrix. It is a busi-
ness event-level modelstorming tool designed to be filled in by/with stakeholders,
using the 7Ws framework. Figure 4-6 shows an event matrix version of the Figure
4-5 manufacturing processes. The additional details on this matrix include:
Event Sequences: Business events, including their main clause short stories,
recorded in time/value/process order.
story telling and
Scrum planning
Dimensions in BEAM ✲ story sequence (who, what, where, why, and how).
This helps you fill in the matrix using the 7Ws, read summary event stories,
spot opportunities to reuse dimensions of the same W-type and focus on con-
forming the most important who and what dimensions: typically customer,
employee and product.
Stakeholder Group columns for recording event interest and ownership. Ticks
can be linked to attendee lists of who was involved in modelstorming the event
details, or should be.
Importance and Estimate rows and columns for prioritizing events and
dimensions on a Scrum product backlog and estimating their ETL tasks for a
sprint backlog.