So all those half-joking "conspiracy theories" about Facebook being a collection tool for three-letter agencies isn't that far off.
I like to think FB is how they check the accuracy of their other data sets.
I don't think tech companies are literally run by three-letter agencies, but after Yahoo and Lavabit (showing two very different reactions to surveillance requests), it's become pretty clear that any large US-based company is at minimum a backup or validation source for surveillance. The standing questions seem to be "do they accept non-warrant surveillance?" and "do they allow bulk surveillance?". If anyone still thinks they don't comply with targeted surveillance, I'd be shocked.
Of course, if you were just getting that information as a matter of course in many arrests and were willing to accumulate it slowly over time it's very possible. I'm preoccupied with scraping issues as that's my only available avenue.