To tell whether a statement about data is over-hyped, see whether it retains its meaning if you replace data with measurements.
So a request like “Please send me the data from your experiment” becomes “Please send me the measurements from your experiment.” Same thing.
But rousing statements about the power of data become banal or even ridiculous. For example, here’s an article from Forbes after substituting measurements for data:
The Hottest Jobs In IT: Training Tomorrow’s Measurements Scientists
If you thought good plumbers and electricians were hard to find, try getting hold of a measurements scientist. The rapid growth of big measurements and analytics for use within businesses has created a huge demand for people capable of extracting knowledge from measurements.
…
Some of the top positions in demand include business intelligence analysts, measurements architects, measurements warehouse analysts and measurements scientists, Reed says. “We believe the demand for measurements expertise will continue to grow as more companies look for ways to capitalize on this information,” he says.
…
I think data is not synonymous to measurement. The meaning of the latter is much narrower and, while I share your dislike of the big data hype, the ridiculousness only stems from this change in scope.
E.g. sales measurements would also sound funny, but just because you don’t acquire this kind of information by measuring something.
@tb: Isn’t it true for data as for measurements, that both are ambigous but data is often associated with objective, unambigous findings? I think that is the point of the substitution game, to be remindet of the volatility of data. You could also replace data with “sampled values”, or “stochastic results”. Data, even without uncertainty of measurement (like counting cars) is only unambigous as a datum. As data, it is just a sample of an unknown distribution.
This is a funny post when you follow it by reading O’Reilly Radar, I’ll never see ‘big data’ the same again.
@tp: I think sales data absolutely is measured, and often depends on many ‘sub’ measurements that specifically involve formal sampling methodologies. I go so far as to suggest that accounting is to sales as statistics is to measurements, with respect to handling the uncertainty inherent in quantifying both.