Thursday, November 11, 2010

Work and play

I don't write about work much, if ever. I like the brainwork I do, but in general a job is a job. That whole career path thing is pretty much only relevant as it pertains to making sure I can provide for my family. Yes I've been asked why I don't change paths entirely and do something else. Well at this stage of the game it would be hard to start over and not take a big drop. However, one thing that I keep considering is combining the aspect of my work that I love with something that could benefit my fellow athletes and friends.

I work in business intelligence and specialize in developing different ways to present information in meaningful manners. One of my strengths is pattern recognition which is used in data mining. Data mining, simply put, is taking large sets of data and looking for patterns in things, possibly previously considered unrelated things, to find new meaningful measures. It's a technique used in a growing area of research termed theory-less research. It's meant to address the all to common practice of developing a theory and then making the data fit that theory rather than objectively determining whether or not your theory holds water. This approach, while a bit difficult at times (there are lots and lots of dead ends!), is intended to more accurately determine causal relationships by ignoring what we believe to be true, which can color the results, and working with only what is true about a given situation.

In the past it seemed that in order to tie work and my passion together meant having to do something coaching related. However after 7 years as a collegiate rowing coach I am pretty much done. That doesn't mean I don't want to help, I just don't really want to be the guy again, not until I "retire." At that point I'll take up my position as a Bill Bowerman-esque running coach, complete with fedora and stopwatch. (I also think coaches should start wearing suits at competitions again, but that's another story for another time!)

But I digress, again. The landscape of training and racing is far different today than in the past. There is sooo much data available. I've had on my personal drawing board a project to combine my data mining skills with this data availability and using some industry standard tools create something really cool in terms of training/racing analysis. I know there are lots of people doing things like this, training peaks comes to mind, but they are limited. They have their "measures" and then market these for everyone. In my world, many of the measures don't exist yet. Oh and there's this, I am really good at this. I trust we'd learn a lot. One area I'm interested in exploring further is the change over time as athletes move from their peaks to their "masters" years. Finding patterns which indicate likely modifications to training protocols would be intriguing. Why can I survive on lower volume than most? I don't subscribe to "genetics" being quite as important as people think. Yes there are differences but I think it's frequently used as an easy out, a cheat. The mapping of the human genome demonstrated one thing clearly, we're all really, really alike.

To date my biggest limiter has been time to do the project and a diverse set of data. Thanks to my Garmin I have lots of data about me. But to do this right I need more than me. Diversity is key. Maybe you're reading this and you have lots of data about you. Well, color me interested. If you'd share I promise to let you share in the the findings. Your name will also be detached from your data (though referenced in any credits), because frankly the interesting stuff is at a higher level than a single person anyway. The possibilities intrigue my brain and my inner athlete. Who knows maybe one day I'll even get something published about this work, something with my picture on it...as long as I'm wearing a fedora.