Posts Tagged ‘The BI Lab’
Introducing…. The BI Lab
OK, so its been a few months since we announced Rittman Mead America and things have been a bit hectic ever since. As I expected, I spent a bunch of time with clients in NYC and the mid-Atlantic. The trip to Singapore was a bit less expected, but it was a great trip and I’d love to go back. Mark just wrapped up some Oracle EPM training there and he seems to feel the same way. I also delivered a couple of presentations at Collaborate and we are coming up quickly on ODTUG Kaleidoscope. Speaking of Collaborate, I would like to thank Peter Reppen for co-presenting one of my talks. Looking forward to Kaleidoscope we are happy to contribute with presentations by Mark and Stewart and by our commercial support as lanyard sponsor.
I’d like to take a little time to introduce myself and share some of my world view on BI. First, I’d like say how much of an honor it is to be able to lead Rittman Mead America, having known Mark and Jon for some time, it is great to be able to work with them as well. I’ve been an independent Oracle consultant for the last 15 years and I’ve been focusing on BI for the about the last 12. I’ve designed and implemented BI solutions for companies in the automotive retail, manufacturing, transportation, commercial real estate, insurance and finance industries.
On the one hand I was one of those unfortunate children who taught himself to program in assembly language in elementary school. On the other hand, I had the opportunity to work with small businesses in my rural town early on. This gave me both a strong enjoyment of technology for technology’s sake, as well as an appreciation for the value or lack of value of technology to a business. These twin and somewhat opposing experiences have defined my view of Business Intelligence. Sometimes as technology people we forget that we are also business people and as BI practitioners we are business people first.
On the business side, the most successful BI projects have the highest level sponsors, if the C-level isn’t involved, its hard to drive value and bring about organizational change. A strategic BI initiative has the ability to change and improve the business and if the sponsor doesn’t have sufficient authority to bring about that change then the opportunity may be lost.
While discussing the business requirements with the sponsor I ask myself what information I would want if I were an investor, owner or purchaser of the company. I have found this leads to questions that help identify and clarify hidden requirements such as information we can easily provide that they didn’t think was attainable.
On the technology side my view of an ideal BI implementation is Linux, the Oracle DB, a well designed Kimball dimensional model, and the current best of breed front end tool (OBIEE). Interestingly since, Oracle first became available for Linux over a decade ago, the only technology piece that seems to changed over time is the front end user tool. I expect this to hold into the future as well.
Finding a space to contribute to such an established, well written, famous blog is something I’ve given a great deal of thought. Having spoken with many readers of the blog over the last few months there are two points I’d like to focus on. Some people seem to have difficulty going about establishing their own lab environment to experiment with and recreate the scenarios in the posts. Another group would like to see how these tools and techniques handle larger data sets.
I’m going to follow the MythBusters methodology, establish a test area, test plausibility, and then over load it and see what happens. Fortunately, the existing postings have already proven the technologies and techniques that work. We can build our test area and then proceed to loading say a billion rows. We will begin by looking at hardware and software options for building your BI lab. Then we will give it a couple of warm up tests before really stress testing it. Along the way we’ll develop a couple of dimensional models and implement them from through the whole technology stack.
