Data Analysis

NYC Taxi Visualizations!

As part of the Great Urban Hack, a team of insanely talented data visualization and transportation geeks did even more crunching on the awesome TLC dataset), and ended up with a closer analysis of what taxi rides look like in the city.

Analysis & Data Visualizations done by Zoe Fraade-Blanar, Kevin Webb, (Fare/Share’s) Aaron Glazer, John Keefe, Lev Steshenko. Data is based on a record of GPS-tagged taxi rides in March 2009 provided by the New York City Taxi and Limousine Commission.



Twenty-four hours of taxi rides (Saturday, March 6th)




Additional Data available in source: Data & Source



Neighborhood Flow Analysis (On an average Monday at 4 pm)


Neighborhood Destination Intensity Maps (Total neighborhood-to-neighborhood flows in a 24 hours period)

Midtown






Upper East Side







Previous data analysis conducted by Fare/Share

General taxi ridership statistics:

  • Almost 7 out of every 10 taxi rides have a single rider, meaning there are at least three empty seats in the taxi – and plenty of room to share.
  • About 1 out of 10 taxi rides are at full rider capacity, indicating there are no free seats and the taxi wouldn’t be available for sharing.
  • The average number of riders per fare was just above 1.5

 

Fare/Share’s Analysis:

We looked at data from more than 14 million taxi rides taken all over New York in March 2009. For that single month, there was an average of approximately 456,000 fares each day, and almost 10 million single person taxi rides.

 

This seems incredibly wasteful — of money, of time, of resources, and of open road. Fare/Share used a geolocation-based analysis to match taxi rides with any similar rides nearby that could have been potentially sharable. We defined a potentially sharable ride is one that has another ride that occurs:

  • Within a three block pick-up radius (0.125 miles)
  • With a ten block drop-off radius (0.5 miles)
  • Up to five minutes before or after original ride’s departure time

 

Using these parameters, we were able to establish that at least 1 out of every 4 current NYC taxi rides could be shared with another rider. So based on those findings, on an average day in March 2009:

  • At least 114,000 taxi rides had potential matches
  • If everyone matched with one other rider, that would be a reduction in over 50,000 cab fares a day from existing fares