[pasbesoin]

Note that this all (certainly, discussion of; I don't know fully about extant realizations) goes beyond what one posts.

There is already analysis arguing (whether or not correctly) that social graphs themselves convey useful, actionable information. For example, the argument that friends of "deadbeats" are likely to be less reliable, themselves. So, if you "friend" too many people having low credit scores, this may effect you when e.g. a lender uses your Facebook graph in evaluating your suitability for a loan.

Similar concerns with regard to evaluation of insurance risk and insurability.

Knowledge asymmetry is the point of arbitrage upon which many people earn a living (and some, a quite extravagant living). Absent sufficient inhibitions (whether self-serving, e.g. Facebook's potential self-destruction; or external, e.g. regulation), people will use "big data" to this effect.