Where to wait for a taxi?

People often have the demand to decide where to wait for a taxi in order to save their time. In this paper, to address this problem, we employ the non-homogeneous Poisson process (NHPP) to model the behavior of vacant taxis. According to the statistics of the parking time of vacant taxis on the roads and the number of the vacant taxis leaving the roads in history, we can estimate the waiting time at different times on road segments. We also propose an approach to make recommendations for potential passengers on where to wait for a taxi based on our estimated waiting time. Then we evaluate our approach through the experiments on simulated passengers and actual trajectories of 12,000 taxis in Beijing. The results show that our estimation is relatively accurate and could be regarded as a reliable upper bound of the waiting time in probability. And our recommendation is a tradeoff between the waiting time and walking distance, which would bring practical assistance to potential passengers. In addition, we develop a mobile application TaxiWaiter on Android OS to help the users wait for taxis based on our approach and historical data.