The taxi that looks for you

Where’s a taxi when you need one? That’s the bane of passengers from around the world, except possibly in Taipei where taxis somehow seem to be just where you need them.

The wait for taxis has spurred the growth of private hire vehicles such as Grab and Uber.

However, such services are not allowed in Japan.

To overcome this challenge, Shin Ishiguro, Data Scientist of NTT DoCoMo , and his team has created an app that helps taxi drivers go where the passengers are with encouraging results for taxi drivers.

NTT DoCoMo uses anonymised data from cell phone signals to detect where people are congregating. It then uses deep learning algorithms running on a NVIDIA DGX-1 supercomputer to map that data to information collected by taxi drivers about demand for their services.

The company delivers information via an app about where passengers can be found to an app that taxi drivers can access from a touchscreen.

To predict demand based on cell phone usage, NTT DoCoMo uses stacked denoising autoencoder algorithms — a relatively new deep learning technique that allows machines to tease patterns out of seemingly random input.

Rolled out to 1,350 taxis in Tokyo and another 1,150 taxis in Nagoya on February 15, the app has steered taxis toward passengers with an accuracy rate of 92.9 percent. That accuracy means drivers in the programme have pocketed 1,409 yen more for every day of driving than taxi drivers who aren’t.