It'd be interesting if comparable data could be gotten for different cities. Wonder what the curves would look like for e.g., Boston, SF, London, Athens, Frankfurt.

Small nit, but guessing this is crow's flight distance which doesn't equal path distance. Regardless, this still isn't surprising. Short rides happen all the time.

:-) this data is equal path as measured by odometer : I was lazy and did not use longitude/latitude for distance. The data itself has miles travelled.

though a bit surprising : 45% of rides are single person rides. But then I wonder : what's surprising : is it low or is it high?

How is the data input on # of passengers? Could it be that it's just easier to hit "1 person" all the time regardless? Are there cabbies that always have the same #?