Taxi data in New York city: A network perspective

Abstract

We work with the "NYC Taxi Data Set," a historical repository of 750 million rides of taxi medallions over a period of four years (2010-2013). This data set provides rich (batch) information on the movements in an urban network as its citizens go about their daily life. We present a spectral analysis of taxi movement based on the graph Fourier transform, which necessitates the spectral decomposition of a large directed, sparse matrix. Important considerations toward handling this matrix are discussed. Preliminary results show that our method allows us to pinpoint locations of co-behavior for traffic in the Manhattan road network.