I think it's very telling that we've had several people telling us almost the same thing, and yet each is contradicting the others.
Apparently, all Big Data questions fit best on Server Fault. No, not Server Fault, DBA. No, not DBA, Super User. No, not Super User, Stats. No, not Stats, GIS. No, not GIS, Stack Overflow ... and so on ...
Here's the thing. Those of us working in Big Data, cover a small bit of each of those areas, and some legal stuff, some organisational psychology, some academia, some web apps, some unix, some ubuntu, some UX, some computer science, and doubtless other areas too. Perhaps someone would like to do a quick check on how many of those Stacks have big-data tag, and how many questions there are on each.
But we only use a corner of each of those fields. We have more in common with other Big Data practitioners, than with specialists in each of those fields.
That's why Big Data is worth a site on its own. And why it's different to just "data".
So I'm optimistic that a Big Data site can thrive; it won't detract significantly from existing Stack Exchanges; it will be open to specialists in all those other fields, who will be able to share their expertise across several Stacks, as many of us already do.