Understanding and visualizing passengers’ travel behaviours: a device-free sensing way leveraging taxi trajectory data

Abstract

Recent years witness the unprecedented proliferation of human-related trajectory data, such as GPS data and Local-Based Social Networks data. Each kind of trajectory data has its pros and cons in instrumenting and charactering human mobility and activities, which provides a valuable opportunity to understand the in-depth intelligence of human travel behaviors from the perspective of data fusion. In this paper, we focus on the motivations of passengers’ travel behaviors, i.e., trip purposes. Rather than deploying sensors to collect data regarding passengers behaviours directly, we propose a device-free sensing way which leverages the taxi trajectory data left when passengers taking taxis. In another word, the data is collected by GPS devices installed in taxis. Specifically, we establish a visual analytics system called VizTripPurpose to understand taxi trip purposes by fusing the taxi trajectory data and human check-in data. Specifically, our system allows for not only obtaining taxi trip purpose on both collective and individual scales, but also visually enable analysts, such as urban planners, to (1) intuitively understand the time-evolving taxi trip purpose patterns by profiling them at different time periods in the multi-scale spatial domain; (2) timely infer the trip purpose for a given taxi trip containing origin, destination, and time information. To illustrate the effectiveness of our system, we provide case studies on real taxi trip data and Foursquare check-in data generated in the region of Manhattan, New York City (NYC), US.