Wednesday, June 26, 2013

Big Data … Big Ideas … Big Thinking

We all keep hearing more and more about how "big data" is changing the world around us...and I am a huge fan!  The opportunity to find patterns and links between things that were previously unknown makes my OCD brain very happy.
But - in the world of Human Resources, I've often found it difficult to articulate why data (big or small) can't always explain the whys of human behavior; e.g. predicting productivity of employees based on some pre-screening test or other personality profiling tool.  I appreciate the implications: "people who scored this way are 75% more likely to xyz", but inevitably, you will be faced with the outlier - the person, or people who did not behave as predicted and how do we explain that?  More importantly, how do we prepare managers and organization to address those circumstances?
I guess ultimately I find some comfort in the fact that humans can be as unpredictable as we are predictable and why I appreciated this posting by David Sable and his ending thought, "Smart people get that algorithms are no substitute for human intuition, insight and judgment." http://www.linkedin.com/today/post/article/20130625162558-234814-big-data-big-ideas-big-thinking?trk=cha-feed-art-title


1 comment:

Unknown said...

I really enjoyed the "for all their brilliance, computers can be thick as a brick" quote.....for some reason it really resonated with me and made me chuckle.

On a more serious note, the part of the article that stuck with me was that "we create 2.5 quintillion bytes of data every day"and that "90% of that data has been created in the last two years." I am not sure what a quintillion is but it sounds like a lot. The real question that this poses is when do we reach the "data convergence?" and will the error term dominate? I define "data convergence" as the point in the future when we have enough information about everything to predict everything with some degree of certainty. Data convergence is not as peachy as it sounds because error, or chaos, will still be present and even more pronounced. The data convergence will simply be a false sense of statistical security.