So this is a good point. I didn't intend for any analytics or inferences to be made with this map, but indeed it seems that density of urban regions is correlated with density of crashes, which isn't too surprising. I wouldn't go as far to say that denser routes are correlated, since the Trans-Atlantic routes are busy but don't have many crashes, as most crashes happen in the terminal environment as opposed to the enroute environment. I could build some analytics tooling on crashes to determine where outliers occur in terms of geographical location (eg. we'd probably expect more crashes in mountainous regions compared to traffic density), so thanks for the idea!