Uber is a ride-sharing application that started as a service for people who couldn't afford a taxi. It operates now in about 70 countries and 900 cities, and it generates over $14 billion revenue.
Uber's team found that sometimes drivers are not close enough when users need them. Uber's research shows that users accept to wait for 5-7 minutes and tend to cancel their ride if the driver takes longer to arrive.
Therefore, Uber's data team would like their app to be able to recommend hot-zones in New York City where the drivies should be.
-
Create an algorithm to find hot-zones in New York City where the drivers should be. The hot-zones should at least be described per day of the week.
-
Visualise the results on a map
I chose to concentrated on data from September 2014 for performance reasons but the approach can be applied to other periods covered by the dataset in analogous manner.