Well, I've violated the most sacred commandment of blogdom: "Thou shalt blog regularly." I won't bore you with lame excuses for why it's been so long since I've posted.
I had the privilege of hosting a Yukon Ascend Business Intelligence (BI) for Developers class in Denmark in December. Reed Jacobsen of Hitachi Consulting taught the class. It's been a long time since I've done any BI projects on SQL Server. The last time I worked on one was for the BizApps conference Microsoft did in Las Vegas back in 1998.
Dave Wickert and I teamed up at that conference to show how to build a procurement data mart on SQL Server 2000. I had the "pleasure" of implementing an extremely sophisticated ETL routine in Data Transformation Services 1.0 that built a data mart from a procurement database, then Dave built some analytics using Analysis Services to crunch the numbers. We brought in some partners to show tool integration during all phases.
At the time we were dealing with version 1.0 (or 1.5) releases of Microsoft's core BI components. I was struck by the possibilities of the tools we were brining to the market. Business Intelligence and Data Warehousing had been such a black art for so long, and finally Microsoft was shipping tools to the masses in every box of SQL Server. With the addition of the Reporting Services late in the SQL Server 2000 cycle, we now had a much more complete offering that added a relational reporting design, execution and management environment.
At the same time I was struck with how far we had to go. The tools were a great first start, but there was still a lot of work that had to be done manually or using custom code. Having taken Reed's class, I finally feel like I have some perspective on what the key advancements are in SQL Server 2005 Business Intelligence.
The world of BI is littered with acronyms and buzzwords. There are religious camps that advocate different methodologies and approaches. This all lends an air of complexity to BI that tends to discourage smart developers from even thinking about it. So the first thing we need to do at Microsoft is to demystify Business Intelligence so that the average smart developer can wrap their mind around it.
First a definition. Charles Fitzgerald once described BI as the following "Unlocking insight from data." Or something approximating that. I love this definition. At its most basic level, BI is about taking raw information and exposing it in ways that allow people at all level's of an organization to make better decisions. Now if you can master .NET, Web Services, SQL, and Visual Studio, you are ready to tackle Business Intelligence.
In my next post I'll drill down on the three core components of Microsoft's BI offering in SQL Server 2005:
Hopefully I'll post sometime before I retire. I the meantime, check out Reed Jacobsen's blog.