So, I am flying home thinking about the event Saturday and many things are running through my head. No, not work on Monday. First, I would like to say that Hadoop is not going to replace SQL Server nor Business Intelligence. Second, a new thing is coming but it is only going to compliment what we already do. And last, Dallas did an excellent job of spreading the learning over 4 tracks with lots to choose from – Visualization, Deployment, Big Data and Analytics. You always had something good to choice from.
The Dimensional Modeling 101 session I presented was great. Lots of sharing of individual experience with creating dimensions and facts. Most of the time was spent on talking about surrogate keys and performance. When I first started doing dimensional modeling, I had the idea that all data came from one transactional system with foreign keys on all tables. The real world proves different. That is what happens after you use some technology for awhile. There was much talk about making mistakes with dimensional modeling and learning from success and failure. There were mostly good comments and 4-5 conversations after the session, during lunch and between other sessions. I look forward to reviewing the critical comments to improve my session abstracts, session content and presentation.
The first session I attended was Intro to Hadoop. The MSBIC group from Dallas was able to get HortonWorks’ speakers which help with explaining a lot about how Big Data works. It is obvious this is a good thing for high output manufacturing systems, logging systems or internet data. Future versions will expand the current features. You could hear many questions about how it all works because you want to wrap your head around the inner workings. The visual display of the Hadoop stack helped tremendously. There is definitely a programming background needed before diving into this new technology. You can download a free sandbox from HortonWorks.
After my session (9:45), I went to a session on pig. It was very hard to follow, so afterwards I went to the speaker room to discuss with others. There, I meet back up with Meagan from Kansas City and Kristen, the leader of the Oklahoma City SQL Server user group. This conversation continued from where we left off from the Friday night speaker social. There is a lot of excitement within the SQL Server community no matter where you are.
The lunch session involved a Microsoft Technical Architect (David Brown) from Dallas that shared his experience with clients in the BI stack. He had great examples that had people laughing and relating to his project management. One comment he made that has stuck in my head is that the problems presented with BI – Self-service, Master Data, Security, Performance, etc. – was “Maybe these problems are good to have for IT.” I think I understand.
The afternoon included a HDInsight session. The cloud is here and you can use it. Next was statistics with R and Mahout which was more about how excited the speaker was rather than demos (that worked). Lastly, was a Slalom consultant talking about Time Intelligence in PowerPivot. He did a good job of listing the functions, but more time should have been spent on simple examples rather than showing everything. The crowd started to ask to many questions that he had to back up and explain.
Finally, I wanted to mention something that happened during the last PASS Summit 2013 in Charlotte that triggered me to come to this SQLSaturday. Karla was talking with me about the Business Analytics conference coming up and I said something about how this was not SQL Server. Her comments made me think back 6-7 years ago when BI started to gain exposure at the Summit and I and others where saying “This is not SQL Server.” Now, I am in the BI stack and love it. I still love DBA work, but I expanded my skills by learning SSIS and SSAS. In Baton Rouge, you need a variety of skills to be employed in the SQL Server realm. You can not just be an expert in SSIS and expect to be employed.
Now might be the time to learn some more skills – Big Data.