Subscribe / Unsubscribe Enewsletters | Login | Register

Pencil Banner

Why you don't get taxis in Singapore when it rains?

Zafar Anjum | Oct. 3, 2012
Big Data throws up a very surprising answer—something that is different from what you might have thought.

"Sometimes data sets can also be combined in new and clever ways. In 2011 Oliver Senn, a senior research engineer with the Singapore-MIT Alliance for Research and Technology (SMART), spent five months working on a joint initiative to give real-time data and insights to citizens to help them improve their city.

"Senn and his team were given the assignment to look at two months of data from a vast taxi fleet of more than 16,000 taxicabs. The massive data set consisted of 830 million GPS records of 80 million taxi trips. But the team had another interesting data set at hand: weather satellite data showing rainfall over Singapore in high resolution. Immediately the direction of the work was clear says Senn, “Everyone in Singapore knows it’s very hard to get a taxi in a rainstorm, so having data about taxis and rain allowed us to look at this”. The initial work resulted in a data visualization shown at an exhibition at the Singapore Art Museum in April 2011 that visually combined the taxi and the rain data, showing what happens to taxis during heavy rain.

"But the study didn’t stop after that. The team had much more compelling ideas that could improve the situation. Once it starts raining in an area, the demand for taxis spikes immediately but then it’s too late. What if a short term forecast of upcoming rain areas could be derived? This information can then be combined with the taxi data and a system could be designed that informs drivers close to upcoming rain areas that they could easily find passengers there - giving them a headstart. Currently the team is working on algorithms that identify drivers close to upcoming areas of rain. The next step is
a prototype system to bring this information to the taxi drivers. Senn sees this project “as a powerful eexample of how data sets can be combined in a new ways to improve the lives of citizens."


Previous Page  1  2 

Sign up for Computerworld eNewsletters.