As someone who has been working in BI for quite a while, i have really enjoyed this take on collecting dw requirements in an agile way, the author is certainly knowledgeable on the topics, and this has given me a structured method to collect requirements. anyone who has been working in BI for a long time can attest to the fact that working in an agile way is vital to delivering a bi project that actually adds value to a business, waterfall is not something that works well with BI
This is the book we've been waiting for in the world of Data Warehousing and Multi Dimensional design. It genuinely tackles Agile for Data Warehousing and does it very well. The secret really is to have iterative collaboration across the board from all interested parties and there are templates and modeling canvases to remind you who should be involved and why. It's a natural and great extension to the wonderful work of Ralph Kimball and augments RB's work and more. I've really enjoyed the very clear notation practice it proposes and it seems to work very well thus far in my practical implementation of the proposed designs. Having been in the DW BI game for 30+ years Ralph's (and colleagues) and now this book makes the job both easier and enjoyable as you are presented with well proven and clear design methodologies. I have personally found it a sheer joy to read with its numerous real life examples and challenges, and better still put into immediate use on a very large scale Data Warehouse project for the Retail industry although it'll fit any industry vertical. Extremely well done Lawrence and Jim.
Corr and Stagnitto have delivered a masterpiece. It's chockfull of dimensional modeling best practices from people who have been there in the trenches, solving the most challenging data modeling problems.
A "model" should make reality look more simple. This holds for data models, or any other model, for that matter. But what does an Entity Relation model typically do? It alienates business people, the stakeholders you need so desperately for a business intelligence project to succeed.
The power of Corr & Stagnitto's BEAM methodology is twofold. First of all, it revolves around language and interactions that business people can relate to: stories and examples. This makes BEAM a truly Agile data modeling approach. And besides enabling a genuinely collaborative and data-driven approach, the output of "Modelstorming" (a pun on brainstorming) is concise, efficient, and can be then be translated to working SQL code with no additional effort. Lean and mean. To this end, Corr has developed a "Modelstormer" Excel template that he has made freely available. By executing the code, you can quickly perform data validation checks on the modeling suggestions from your business partners. It doesn't get any better, or more Agile!
Besides the BEAM methodology, the book goes into considerable detail on how to deal with some particularly tricky dimensional modeling challenges. Ragged hierarchies, recursive relationships, hybrid SCD's, and several design patterns that are notoriously difficult and interesting to optimize, form the last chapters of the book.
This book should be mandatory reading for any serious (dimensional) data modeler and information analyst in the BI space!
I love Business Intelligence and data warehousing, in fact you could say my career is my 'hobby'. There are very few books I have read on the topic in recent years that have excited me as much as this one! It confirms what I always suspected - data warehousing CAN be done in an agile way!
The BEAM* approach as laid out in this book is so 'accessible' and practical to EVERYone who may be involved in a data warehousing project, from business users, to business analysts, project managers and even the most hardened techy developers. It creates a new 'easy language' for communication between these different contributors to the project/solution. The method is inclusive and collaborative, as well as thorough and logical, yet so simple and practical, and can be adapted to any complexity with ease i.e. agile.
The book it's self is nicely laid out, with plenty of real examples to demonstrate the concepts. Even concepts I was familiar with before, are explained very well and a great intro for anyone not familiar with or a bit daunted by data warehousing.
I recommended this book to my colleagues (PM's, BA's, developers, architects) and since going on the course with Lawrence they haven't stopped talking about the methods and can't wait to unleash their new found skills on the next analaysis meetings we are having with the business.
I personally think the approach is 'revolutionary' in it's simple elegance and real life usefulness, and will be recommending this book with all my BI passion to anyone remotely attached to any data warehouse project!
A superb book that presents a powerful, and yet easy to use, approach to capturing business requirements for data warehousing.
The BEAM* method (business event analysis and modelling) beautifully updates the discipline of analytical data modelling using contemporary agile principles, and does so in an easy manner which I found to be a stark contrats from the somewhat dry text I have read on the subject to date. It is clear from chapter one that the authors have a vast experience and understanding of the field of data modelling includes OLTP as well as OLAP systems.
I have worked in the BI & Data Warehousing space for over 10 years and found answers to a number of situations, such as how to handle a column whose SCD type varies according to changes in other dimension columns, which were not easily forthcoming even during my time working for one of the top global data management and analytical software companies.
This book is simply a must for modellers, ETL developers, and importantly business persons, who are involved in the process of requirements capture and model generation. The approach accurately captures the current needs, in a timely manner, that can be easily understood by both business and IT professionals alike. This is the solid foundation which should lead to the production of a highly useful data warehouse that genuinely meets organisational needs.
I bought this book when I noticed the high number of 5 star ratings. I have read it cover to cover and YES it does deserve all 5 stars. The first half of the book provides a simple framework for requirements gathering and modeling a data warehouse with business users. The book really gets technically interesting in the second half when the author explains each dimension type in detail with it's typical business intelligence problems and gives detailed design pattern solutions for modelling them. I really liked the techniques on optimizing large FACT tables. It is very well written, and also provides hints on actual SQL for building and using the dimensional models. I find it very close to actual business scenarios rather than a high-level strategy book about data warehousing. It is always on my desk and I keep referring to it at my work place. The only suggestion I can give the author is to provide an e-book version soon for quick anytime access. Money well spent!!
This is the guide/handbook for the BEAM methodology in which Lawrence lays out a truly collaborative and agile approach to the design of Data Warehouse Dimensional Models.
Whilst agile approaches have been successfully applied to the delivery (ETL/coding) phases of Data Warehouse projects a methodology to enable agile design has long been needed, BEAM delivers this.
The book begins at first principles however a decent knowledge of database and Dimensional Modelling design techniques will certainly ease your learning curve. The methodology then provides a comprehensive framework to enable the interactive and collaborative design of dimensionally modelled Star Schemas and also includes a number of very useful common design patterns.
All in all, highly recommended reading for anyone in the Dimensional Modelling Data Warehouse design space.
For any experienced BI and Agile BI practitioners the constant challenge is how to get sensible, meaningful and deliverable requirements from business users quickly and easily. Having tried this approach a couple of times BEAM has the answer! BEAM is an excellent way of breaking down the requirements into simple, realistic tasks that do not over-tax the brain-cells of users. BEAM is well suited to an Agile approach that adheres to the Kimball design methodology.
At last, a book that knows what it is talking about, not just a lot of regurgitated BI flummery. This tool neatly avoids the dread catch-all requirement cliches of the "busy user". i.e. - no more "just rebuild all my reports", - no more "I don't know what I want, show me something and I will tell you what's wrong" - no more "300 page requirement specs to review.
This book describes a very pragmatic approach for designing data models for data warehouses.
Agile methods and data warehouse design meet each other here. Very clearly written, no fuss, and focused on keeping it simple.
It takes the Fact Dimension Diagram (FDD) to a whole new level, and accompanies it with very well worked out examples on how to structure the design process and eventually get a complete design, based on the idea you do not need to understand business processes in detail beforehand, but Just Enough to deliver functionally working data warehouse evolution after evolution.
If I would have to list some subjects I think differently about it would be:
- the data warehouse matrix: I would maintain details in a data modelling tool and use extended attributes and scripting to generate the data warehouse matrix from it; my point is that the data warehouse matrix is very much related to the physical data model and maintaining the data warehouse matrix AND a data model seems to go against the pragmatism of the rest of the book (the book does hint on using extended attributes in data modelling tools, and a lot of details concern standard facailities within data modelling tools, such as null values, min, max, primary key, etc.)
- the author writes that roll-up dimensions should be bases on the base dimension; I disagree, because the roll-up dimension could list more information than could be derived from the base dimension; for example, there can be more product types in use and needed for conformed dimensions than could be derived from one of the base product dimensions; always use the source in my opinion