number of deaths from autos hitting bicyclists per capita
I think some other ideas have alluded to this, ie attitude of drivers towards bikers, awareness, etc but what are your chances of simply riding around a town or area and getting hit by some idiot driver? I live in San Antonio, and while a pretty weather friendly, relatively flat town with allot of bikers, there has been quite a few auto bike related deaths in the last year and I think this needs to be called out for calculating a bike score.
This would have to be normalized as per bicyclist, not per capita. If lots of folks in a city commute by bike, bike accidents per capita will likely be higher than a city where no one bikes. The trouble is, it's hard to get accurate numbers of who bikes and how much. Some data exist from relatively wealthy enthusiasts who log and share their gps cycling information, but that's a rather narrow slice of the cyclists you see on the road.