There are datasets that allow for inference on social interactions in various context. BoardEx for example will allow you to figure out who's been hanging out with whom in golf clubs for board members of major companies. This is as un-open data as it could be, however, some universities have licences so you might be able to get access.
Also twitter data would also be publicly available and easily accessible through the Twitter API.
Beyond the data issue, you are facing serious identification issues. How will you argue causality? How will you argue that any correlation you observe is not just the result of e.g. better looking people earning more and also being more socially active? However, I don't think you'd need to convince anyone that having a specific kind of connections is generally good for your career, all those networking sites/events/agencies serve as evidence for that. And the literature on that is huge, just start reading up on social capital and you'll find an endless literature from sociology, economics an related disciplines.
I am not sure where you are heading with this (scope-wise), but if this is about a thesis or so you might want to narrow yourself down substantially...