It’s not so much about the sites that house the data, although it is in the way it collects them. You see, there are so many sites sprouting up randomly each day, with intermittent users, successes and failures, with tons of data being created.
It’s all about that data.
But not just the raw data, but the way it’s assembled.
The ability to “export” this data, move it around, create interesting sets with it, is really what its all about. This data can give us insight into new and exciting fields, even if that data was created on the premise of something entirely different.
Take Facebook for example. It is simple so big, it generates a ridiculous amount of data on a regular basis. And since its a social site revolving around NOTHING in particular, it generates loads of raw social data. This information can be used across many spectrums, from understanding relationship trends to the myriad concepts of game theory behavior.
So in itself, Facebook is… not so useful. But the data it generates is very useful.