Can Data Take the Place of Hypothesis?

For years, folks have been talking about the analysis of huge data sets (and by huge, I mean really, really huge) superseding our ability to hypothesize.  A hypothesis to prove or disprove, that essential building block of the scientific method, is becoming less important than just knowing everything you can know about the problem.

This is the best example I have seen so far that this might be the case.  Deb Roy's analysis of his son's language acquisition yielded several simultaneous findings because the data was so rich and there was just so damned much of it.  Sure, you have to decide which axes to measure.  But what makes massive data manipulation so interesting is you don't necessarily need to know why you're interested in those axes.  Science becomes a more improvisational, reacting to the data-shapes changing right in front of you.

You still have to see the results, you just don't have to predict them quite as much.