Ask HN: How can someone best improve Wright's data?

Recently, Tony Wright (webwright) compiled some quick data on startup acquisitions to hopefully to get a picture of the geographic distribution of startups. It was quick, dirty, and easy to poke holes into, but enough to consider some further investigation. So, what would be the best way to gather information on startups while minimizing bias? I'd be willing to put some money into Mechanical Turk to do the legwork if there's a halfway decent method to it.