Boston Taxi Pickups in D3

Screen Shot 2016-05-20 at 8.34.44 PM


As part of a homework assignment for a class on data mining, I got to explore a dataset of taxicab pickups in Boston.  With 6 months of data and the geocoded locations of roughly 4.2 million cabs, our task was to show how taxi demand differed across  attractions around the city.  To make this visualization I laid a 200m by 200m grid over the city, they counted the number of taxi pickups in each grid cell for each hour of the week (all weeks were summed together).  I then identified the grid cell containing each of the 36 attractions and plotted the time series using D3.js.

I had expected that each time series would look roughly the same, displaying the circadian rhythm of human behavior, but there is surprising variety in the data without any sophisticated statistics.  I particularly like the time series of different theaters which seem to peak Thursday, Friday, and Saturday nights.  Also interesting is the contrast between the 27 Drydock and the Institute of Contemporary Art.  One calls for cabs late at night, the other very early in the morning (maybe night shift workers?).

The code can be found here:


Leave a Reply