My first Publication Agile-Data-Warehouse-Design-eBook | Page 16

I NTRODUCTION
Dimensional modeling, since it was first popularized by Ralph Kimball in the mid-
1990s, has become the accepted (data modeling) technique for designing the high
performance data warehouses that underpin the success of today’s business intelli-
gence (BI) applications. Yet, with an ever increasing number of BI initiatives
stumbling long before they reach the data modeling phase, it has become clear that
Data Warehousing/Business Intelligence (DW/BI) needs new techniques that can
revolutionize BI requirements analysis in the same way that dimensional modeling
has revolutionized BI database design. Dimensional
Agile, with its mantra of creating business value through the early and frequent
delivery of working software and responding to change, has had just such a revolu-
tionary effect on the world of application development. Can it take on the chal-
lenges of DW/BI? Agile’s emphasis on collaboration and incremental development
coupled with techniques such as Scrum and User Stories, will certainly improve BI
application development—once a data warehouse is in place. But to truly have an
impact on DW/BI, agile must also address data warehouse design itself. Unfortu-
nately, the agile approaches that have emerged, so far, are vague and non-
prescriptive in this one key area. For agile BI to be more than a marketing reboot of
business-as-usual business intelligence, it must be agile DW/BI and we, DW/BI
professionals, must do what every true agilist would recommend: adapt agile to
meet our needs while still upholding its values and principlFs (see Appendix A: The
Agile Manifesto). At the same time, agilists coming afresh to DW/BI, for their part,
must learn our hard-won data lessons. Agile techniques
With that aim in mind, this book introduces BEAM ✲ (Business Event Analysis &
Modeling): a set of collaborative techniques for modelstorming BI data require-
ments and translating them into dimensional models on an agile timescale. We call
the BEAM ✲ approach “modelstorming” because it combines data modeling and
brainstorming techniques for rapidly creating inclusive, understandable models
that fully engage BI stakeholders. This book is about
BEAM ✲ modelers achieve this by asking stakeholders to tell data stories, using the
7W dimensional types—who, what, when, where, how many, why, and how—to
describe the business events they need to measure. BEAM ✲ models support mod-
elstorming by differing radically from conventional entity-relationship (ER) based
models. BEAM ✲ uses tabular notation and example data stories to define business
events in a format that is instantly recognizable to spreadsheet-literate BI
stakeholders, yet easily translated into atomic-detailed star schemas. By doing so,
BEAM ✲ bridges the business-IT gap, creates consensus on data definitions and
generates a sense of business ownership and pride in the resulting database design. BEAM✲ is used for
modeling is
responsible for
today’s DW/BI
successes, yet we
still struggle to
deliver enough BI
can help, but they
must address data
warehouse design,
not just BI
application
development
BEAM✲: an agile
approach to
dimensional
modeling
modelstorming BI
requirements
directly with BI
stakeholders
XVII