Wednesday I attended the second annual Data Science Summit, hosted by Greenplum. The general topic of data science has intrigued me for quite some time, but I've had very few chances to talk to people actually in the field -- i.e., those who are turning theory into practice. After all, many cash-strapped organizations are struggling to keep the lights on. Not all that many are venturing into mostly unchartered territory.
At breakfast, I spoke with Scott Kahler of AdKnowledge, John Paclik of Comcast, and Aaron Caldiero of Zions Bancorporation about some of the very specific (read: technical) lessons learned in deploying tools like Hadoop. To be sure, this was not a discussion for the technically challenged. I was glad that I could keep up and not have to have every acronym explained to me. Turns out that Big Data technologies aren't entirely unlike more mature ones like ERP and CRM: there are tradeoffs to be made and one size rarely fits all. The benefits of Big Data, though, are hard to overstate.
The panelists and featured speakers then took over. While all interesting, I particularly enjoyed Oren Etzioni of Decide.com who talked about his company's price comparison site and app. Sure, price comparison sites like mysimon.com (great name, right?) have existed for a long time. However, Decide uses Big Data to do something truly unique: it helps you decide when to buy something. Etzioni is a fellow Carnegie Mellon alumnus and ridiculously smart cookie and it will be interesting to see how his company's offerings evolve. (Interesting factoid: the prices of large-screen TVs don't always decline. They sometimes actually creep up.)
Big Data is ridiculously powerful and sometimes scary. Exhibit A: Target evidently knows its teenage customers better in some cases than their parents do. Privacy generally seems to take a back seat these days, but at least most of the speakers gave lip service to that quaint old notion.
On a different level, the number of Moneyball references really shocked me. Perhaps CIOs are finally geting analytics. After all, if it worked for the Oakland A's, then perhaps it can work for their companies?
What say you?