New Popular Science Books discussion

Just finished reading the new book by Viktor Mayer-Schonberger and Kenneth Cukier called Big Data: A Revolution That Will Transform How We Live, Work, and Think. The book traces the history, present and future of big data, and argues that the phenomenon will transform our world. It's a good overview of an interesting and important topic. I've written a full executive summary available here: http://newbooksinbrief.com/2013/03/21...

Cheers,
Aaron

David Brooks writes about the book today and seems to disagree with the hands-off attitude toward the data. Or maybe I should say "minds-off". He would prefer that we do think about causality and not just trust the numbers.

http://www.nytimes.com/2013/04/16/opi...

Peter wrote: "David Brooks writes about the book today and seems to disagree with the hands-off attitude toward the data. Or maybe I should say "minds-off". He would prefer that we do think about causality and n..."

Yes. Nate Silver expresses this opinion as well. It's always going to be dangerous to assume that a correlation indicates causation, since this simply isn't always the case. The authors of Big Data, at times, seem to recognize this as well, but ultimately they shunt this concern aside. Somewhat irresponsibly, I think.

I guess it depends how much you need to understand what you do. Some models, for instance, don't need to understand just why smiling puts money in their bank account and they would be foolish to look for a serious understanding. But I think business is a bit different. It's dangerous to invest too much in plan just because some model showed some correlation. It might be something temporary that's going to disappear and it sure helps to have a theory to back up the investment.