These are miles driven in each location, not miles driven by drivers from the location, correct? So when Colorado drivers travel through Wyoming, that would count as miles for Wyoming? Assuming so, states with small populations and a high fraction of out-of-state drivers traveling through will rank disproportionately high. Wyoming likely fits that description. States with high populations and fewer out-of-state drivers will rank lower. Washington DC and New York certainly do have a lot of out-of-town visitors, but those visitors probably aren't racking up a ton of driving miles in those places, given the urban settings and plentiful transit alternatives. It would require a completely different data collection mechanism, but I'd find it more interesting to see these results grouped by drivers' states of residence rather than by where the miles are being driven.