I decided to monkey around with the criteria, asking myself: what if I just used transit ridership (or more broadly, non-auto commuting) without using the poverty rate as a factor? Does that change my rankings?
In particular, I created grades as follows: over 60 percent non-auto (defined as “transit plus walking” since the Census doesn’t ask about bikes) = A, 50-60 percent = A-, 40-50= B+, 30-40 = B, 25-30 = B- , 20-25 = C+, 15-20= C, 10-15=C-, 7-10= D+, 5-7=D, 3-5=D-, below 3=F.
I do not think any city’s grade changed by more than a point and a half. However, some cities were so closely bunched together that it mattered.
In the “cost of living weighed as 30 percent” category, Seattle and Minneapolis were still the leaders- Seattle winning with 84, Minneapolis a close second at 83.5. Salt Lake City was third at 83, and then Providence, St Paul, Pittsburgh, Rochester and Denver were all pretty closely bunched together. Birmingham continued to be dead last, with Miami and New Orleans being the only other cities below 75.
In the “ignore cost of living” category, the changed rankings put NYC ahead of SF (90.4 to 89.5), then Boston (87.6), Seattle (86), Minneapolis (84). Birmingham was again last, with St.Louis, Indianapolis, New Orleans, Kansas City and Nashville all pretty closely bunched together.