Measuring the Pulse of a City Via Taxi Operation: Case Study

Taxis are vital to modern urban transportation systems. As a great complement to public transportation, taxi services are popular for their flexibility and availability. The ever-increasing magnitude of taxi services has generated a large amount of data about vehicle locations and trips, making it possible to investigate taxi operations in detail. Previous work on this topic has addressed aspects of taxi services such as trip type clustering, visualization of taxi trips and taxi dispatching mechanisms. Based on a data set of taxi trips in New York City, in this paper we analyze a taxi system from some new perspectives. First, from records of fares collected, we investigate driver and passenger behaviors; specifically, what makes some drivers successful and how passengers tip. Second, we study the trips themselves, finding a strong periodicity in the data; for example, the number of pick-ups, the average speed and the average tip ratio are highly periodic in a location-specific manner. This periodicity suggests that each neighborhood in a city has its own “pulse” which makes the taxi system of a city behave in a predictable fashion. Third, we examine the impact of various socio-cultural factors on taxi operations. For example, the average tip ratio of taxi trips can be affected by the outcome of a sports game; fans tip more when their teams win. These results indicate a close connection between taxi systems and human activities, further connecting the pulse of a city to the operation of its taxi systems.