Big Data for Boston: to improve transport, city enlists MIT students to crunch numbers

Bookmark and Share

testBy Adam Conner-Simons

What happens when you give a bunch of MIT students GPS information on more than 2.3 million Boston taxi rides, and then offer them a cash prize to get creative with the data?

That was the idea behind the inaugural MIT Big Data Challenge organized by the MIT Big Data Initiative at CSAIL in partnership with the City of Boston and Transportation@MIT.

With urban congestion on the rise, planners have been looking for new ways to understand and improve transportation in Boston, which spurred the decision to collaborate with MIT and make available data from local events, Tweets, weather records and taxi rides in commercial zones of the city.

“We believe that big data can be used in the service of big issues,” said Elizabeth Bruce, Director of the MIT Big Data Initiative at CSAIL. “We thought it would be exciting to organize a challenge in which our community of students and researchers put their energies into tackling a tangible public-policy issue like transportation, and to see what insights emerge from that."

Since November, more than 250 teams submitted work focused on trying to predict demand for taxis and create intuitive visualizations about these topics. Among the questions that teams explored:

At the final event February 27, teams presented their work and heard from a panel of judges that included the City of Boston’s director of transportation planning, Vineet Gupta.
 
“We are excited about this partnership with MIT because it opens up a whole host of opportunities for the city,” said Gupta. “The data analysis that these students have done reflects true out-of-the-box thinking, and has the potential to directly inform future policy decisions in Boston.”
 
Teams took first prizes for prediction and visualizations, respectively. The prediction winner was Suma Desu, a master's student at the Human Mobility and Networks Lab. The visualization winner was CSAIL graduate student Gartheeban Ganeshapillai, who created an interactive map that allows users to see how taxi patterns change over time. He also built a visualization that displays the most common intracity rides. (The winner, with over 10,000 rides, is the trip between Hynes Convention Center and the Cutler Majestic Theatre.)
 

See below for a list of the top three finalists as well as some fun facts taken from the data. The bigdata@CSAIL site also has a gallery of all visualization submissions and a complete list of winners.

Taxi data was drawn from 2.3 million rides that occurred between May and November of 2012. Data partners for the event included Creative CMT, GNIP, MBTA, Telenav and Twitter.   

VISUALIZATION
FIRST PLACE - "Team GARTHEE" (Gartheeban Ganeshapillai - MIT CSAIL)
SECOND PLACE - "Team JAMESON" (Jameson Toole - MIT Human Mobility and Networks Lab, Engineering Systems Division)
HONORABLE MENTION - "Team KVGH" (Kael Greco, Veronica Adelle Hannan - MIT)

PREDICTION
FIRST PLACE - “Team HUMNET” (Lauren Alexander, Serdar Colaxk, Suma Desu, Jameson Toole, Yingxiang Yang - MIT Department of Civil and Environmental Engineering)
SECOND PLACE - "Team PIGGY XU" (Runmin Xu - MIT Department of Civil and Environmental Engineering)
THIRD PLACE - “Team MATTED” (Matthew Edwards - MIT EECS & CSAIL)

Boston's 10 most popular taxi pick-up stops
1. Hynes Convention Center, 76.1K rides
2. Wilbur Theater, 51.4K
3. Majestic Theater, 51.3K
4. South Station, 51.1K
5. Shubert Theater, 49.4K
6. Wang Theater, 47.4K
7. Faneuil Hall, 45K
8. Charles Playhouse, 41.2K
9. Post Office Square, 37.7K  
10. Back Bay T, 34.6K
(Based on Team GARTHEE's Intracity Rides visualization)

5 cab rides that would be faster by bike
Hynes Convention Center to Back Bay T
South Station to MGH
Boston Common to Prudential Center
Aquarium to World Trade Center
MGH to North End
(Based on Team STETNER's research)