Tuesday, November 12, 2013

MIT Big Data Challenge: Predict Taxi Demand

The goal is to predict taxi demand in downtown Boston based on historical taxi data (2.3 million rides) and other related data sets (weather, transit, events, and social media data).
"We have collected several months worth of twitter, MBTA, taxi pickup and drop off, and other datasets (see datasets page). In this challenge we will give you all of the data except for a number of 2 hour time spans at several locations in downtown Boston that we will remove from the taxi pickup dataset. Your job is to predict the total number of taxi trips within 250 meters of the locations and the 2 hour time spans. We will give you a list of location (latitude, longitude points) and timespan pairs. Your score will be computed as the root-mean-squared error of your predictions against the ground truth."

No comments:

Post a Comment