Is Data Vault Agile?
You bet it is!
Years ago I wrote an article about Agile Data Warehousing and proposed using Data Vault Data Modeling as a way to get there. Dan Linstedt recently published an article with more details on how it fits. Here are the good parts:
1. Individuals and Interactions over processes and tools
The Data Vault is technology agnostic AND focuses VERY heavily on customer interaction. In fact it’s really the only methodology where I’ve seen a very heavy emphasis on the fact – The business owns the data.
Also, you have to start with the Hub entities and they require identification of the business keys as specified step-by-step on page 54 of Super Charge Your Data Warehouse
2. Working Software over Comprehensive Documentation
With the pattern based architecture in a Data Vault model and with the business rules downstream of the Data Warehouse, you can start building extremely fast and even use code-generation tools or scripts to get the first cut of your model.
I’ve in fact used code-generation for Data Warehouses that have been in production for quite a few years They’re even running today.
The Data Vault Model & Methodology in my opinion is the fastest way to get something delivered to a Data Warehouse and it dramatically reduces project timelines and risk.
3. Customer Collaboration over Contract Negotiation
The Data Vault Methodology emphasizes the ownership of the project and data by the business and makes them a partner on any Business Intelligence project.
And, the fact that it’s easy to address change makes them happy which interestingly enough, is the next one:
4. Responding to Change over Following a Plan
This makes some people cringe. But it’s a reality of most projects. The first time out neither you nor the business REALLY know what they want. It’s only after they see something, they realize the value of the information and their brains start churning.
In the traditional forms of Data Warehousing, it takes scoping, project budgeting, planning, resource allocation and other fun stuff before you can even get creative and give them what they think they want. The problem is, most business users don’t REALLY know. The DW team ends up thinking and even assuming for them often incorrectly. You can end up with something that is really fancy and beautiful and still … useless!
To add to the complication, If it’s in fact a bad idea, it will be money ill spent which can be as much of a big deal if it’s a great idea where the time to build will make them lose out on the competitive edge they’re looking for.
With the Data Vault, the model is built for change from the ground up. Since the core data NEVER ever changes, creating business level user-interface layers on top is just so easy – and many architects and modelers think it’s ideal.
Check out the full post – Agile Data Warehousing
(and don’t forget to buy the book).
BTW – if you are going to ODTUG KScope12 this June in San Antonio, be sure to stop by for a chat. I will be giving two talks, one on Data Vault and one on using SQL Developer Data Modeler.
P.S. I am now on twitter! You can follow me there @KentGraziano.