That's fair; props on all of the effort you've put into that! I'd imagine one of the toughest challenges is subtracting the volatile "layman demand" and gamblers to get to that fundamental value derived from actual users.
The nice thing about indexing the whole web is that you also have a complete per-author picture of everything they've written. Shills often post the exact same message on many different sites, or talk exclusively about one project and nothing else. Gamblers & laymen tend to pop up abruptly when a project gets hot, and then quickly either disappear or move on to the next hot thing. When a veteran user with a reputation for detailed, often-quoted posts starts talking about a particular project, though, and keeps talking about it, usually there's something real there.