Stephen O’Grady explains why he’s taking a statistics class:
Life, according to economics, is about incentives. My incentive to learn such things is simple: the ability to be able to understand more completely what data is trying to tell us will have value. Value more than sufficient to offset my investment. Or so I hope.
I would love to learn more about statistics for a very simple reason — so many of the things that interest me most these days were written by people who are using statistical methods to break down data. Whether the topic is sports, economics, or political science, people are using statistics to look at old problems in new and interesting ways.
I’m also seeing more and more ways that a better understanding of statistics could make me better at my job. In software development, we’re a lot better at gathering data than we are at analyzing that data to turn it into useful information. In many cases, we look at performance information and have a hard time distinguishing between noise and clues. Getting better at that requires deeper math.
April 8, 2010 at 8:25 am
I was just thinking about this same exact issue just yesterday in a marketing meeting. I always knew statistics was an area I could/should brush up on, but sitting down with someone who really knew how to analyze data pretty much blew me away. He was able to coax information out of all those numbers that none of the devs had even considered. It was like realizing you’d been driving around in second gear for the last six months when gears 3-5 were just waiting for you to use them.
This isn’t to suggest that we should all be thinking like marketers, but I definitely believe that your better developers and the team leaders could benefit from the ability to get more out of all that data we have access to.
April 8, 2010 at 10:31 am
You post brought to mind this post: http://www.zedshaw.com/essays/programmer_stats.html
April 8, 2010 at 12:36 pm
A somewhat-related observation. I find that organizations are great about having methods to record data on projects but the tools to pull meaning from that data are lacking.
Just as there’s a lot of performance data to analyze, I’ve seen that a lot of “executive summary” information is already collected, but the means of digging into it is very time consuming.