Agile Data Warehouse Design: Collaborative Dimensional Modeling, from Whiteboard to Star Schema

When it comes to dimensional modeling and related data warehouse and business intelligence (DW/BI) topics, until now the "Big Two" authors who provide truly useful books are Ralph Kimball (along with his Kimball Group colleagues) and Christopher Adamson. With the publication of Agile Data Warehouse Design: Collaborative Dimensional Modeling, from Whiteboard to Star Schema (ADWD), Lawrence Corr (along with his co-author Jim Stagnitto) contributes critical new information and techniques to the field - and we now have the "Big Three" dimensional design authors. Why are these my "Big Three" BI/DW authors, and why does Lawrence Corr deserve to join the list? The numerous Kimball Group books and the three Adamson books all provide a useful blend of theoretical and practical information and techniques to guide the DW/BI professional in creating systems that are affordable, maintainable, accurate, and useful to clients. In his new book, Corr fills in a critical missing component: how to effectively engage business and technical personnel in a dialogue to perform agile design where the emphasis is on content, understandability, rigor, and usefulness, instead of on producing reams of documentation. In short, this book is about agile dimensional design including "why" and the all-important "how to" with supporting templates. Though many of the book's topics are previously addressed by other authors, Corr expands upon these and presents a large set of fresh ideas and techniques that are robust and adaptable, while incorporating the concepts of the Agile Manifesto. He provides an emphasis on collaboration and flexibility, not just on theoretic agility, while adapting a method called BEAM* (Business Event Analysis & Modeling) for dimensional design. BEAM* is an agile modeling method that lends itself to rapid, collaborative dimensional design sessions with business and technical participants. BEAM* utilizes a set of diagram types that, taken together, provide a complete design that is understandable by business people and is immediately useful for implementation people. The diagram types are: 1. Example Data Table (or BEAM* Table) - Primary diagram type used to capture data stories and describe data requirements through sample data. Supports modeling by example rather than by abstraction, making them useful to all participants in the design activities. Uses sets of short codes to capture and indicate design details. 2. Hierarchy Chart - Shows hierarchical relationships among related entities within a dimension. This is a much cleaner and more concise version of the traditional dimensional hierarchy chart. 3. Timeline - Concisely captures sequences and durations as an aid to understanding and design. 4. Event Matrix - A fresh twist on the classic fact/dimension matrix, with better organization, which provides an accessible overview of multiple star data mart and data warehouse designs. 5. Enhanced Star Schema - Traditional star schema diagrams augmented with BEAM* codes to indicate dimensional design aspects that are not supported by traditional modeling tools. These "tools" facilitate design discussions, detailed design work, communication, implementation, and provide concise yet through documentation. In addition to the methodology and diagram types, extensive coverage of the design process and the designs themselves are covered. All of the dimensional modeling basics receive thorough treatment and many of the more challenging dimensional modeling problems receive substantial coverage with mature techniques and solutions drawn from Corr and Stagnitto's extensive consulting work. Examples include several types of hierarchies and hierarchy maps, multiple calendars, time (clock time), and the always exciting Customer and Employee dimensions, among many other design challenges. Agile Data Warehouse Design is an eminently useful book and a long-needed complement to the dimensional modeling literature.

I admit that I was a bit jaded about the true effectiveness of Agile and EDW (after all everyone seems to be jumping on the Agile bandwagon and using it as a buzzword). However, after reading this book, I now do see how Agile and EDW could work. It is not a silver bullet for eliminating re-coding or designing a data warehouse that will cover all future reporting needs but it should help in developing a robust, proactive design that anticipates future demand and designs for it. The BEAM* methodology alone was worth the "price of admission". It provides a manner to solicit and capture requirements (using Agile approaches of course) that makes is easier for stakeholders (people who will run and request reports) to articulate requirements. And then, a manner to translate them into Dimensional Models.

Good book overall, obviously not a page turner. I thought this book was a little too vague for me, but really gave me a nice overview of the process and hopeful end result. While the book was easy to follow, the design choices used in it were a little old-fashioned and seemed to make it look like a cheap at-home printed textbook.

I was a Ralph Kimball fan since the late 1990s. In 2012 I came across Lawrence's book. It clarified the process and dimensional design patterns very clearly. In addition, Lawrence was generous about sharing his materials and templates and providing ideas when I contacted him. If you are a BI professional who works directly with business stakeholders and who wants to collaborate effectively, this is the book for you. If you are a dimensional database modeler who wants to know more about dimensional design patterns and when to use them to provide the best product for your users, read this book. If you are a BI professional who simply wants to know what is possible in terms of a flexible, efficient and extensible design, read this book.

I have worked in this industry for over 20 years and this is my favorite book on Data Warehouse design. The author has done a great job of "trimming the fat off the meat" to keep the focus on the most valuable content. For practitioners, this book is a great reference to quickly find the information you need to craft an appropriate solution. The book is well organized with great notes in the margins and very helpful chapter summaries. This book is packed with relevant examples to clearly illustrate the concepts. In 2001, I took the author's course and have been referencing his course materials for years and can highly recommend this book! Entry level to expert, this is your book!

Far and away the most comprehensive, pragmatic, and insightful book on warehouse design I've ever come across. In a dense but approachable (and beautifully written) 300 pages, this book recapitulates and re-examines the accepted canon of work in this space (Kimball, et al), reinvigorating the well-worn fundamentals of data warehouse model design; it's a primer and an expert-level guide in one. Ideas and techniques that have only previously been described in long-winded and subjective prose, are re-examined and made precise and applicable. Applying the core notions from Scott Ambler's Agile Modeling, Gojko Adzic's Specification By Example, and Mike Cohn's User Stories Applied, along with the best of everything Kimball, et al, have ever had to say, the BEAM methodology guides and organizes the practitioner's thinking, providing the recipes and patterns necessary to consistently deliver quality models.

First industry specific book that I felt was not a forced read for me. I've read large parts of the Kimball Desk Reference set :) and various others on the subject of modeling and managing projects of the BI/DW focus, but each felt like a text book compared to how well this has been written and distilled. I've had success in the past with a mentor that has insisted using spreadsheets to tease out data models for OLTP systems, and have found greater success in communicating and capturing designs with less technical audiences. Congratulations on formalizing that approach when Modeling Business Events! I immediately see how this will benefit the entire life-cycle.

This may be an effective approach for some groups, but was too much framework overhead for my tastes. In my opinion, this kind of highlights the discrepancy between the original agile manifesto and the agile methodology. I'd have preferred paired down content that focused on the core concepts.