Regularity and predictability of individual spatiotemporal behavior: A case study of residents working day activities in shanghai

Abstract

Understanding the regularity and randomness of residents' daily activities is fundamental for spatial behavior studies when the focus is turned from description to prediction. In this study, behavioral regularity was explored from multiple aspects based on individual spatiotemporal trajectories, which were extracted from 10 working days' mobile signaling data in Shanghai Municipality. First, spatiotemporal behavior was recognized as four modes—non-trips, random trips only, regular commutes, and other regular trips—based on periodical recurrency and time series connection. The latter two modes with strong regularity account for 49.3% in population, 83.6% in trip generations, and 82.0% in trip distance. Second, entropies of trajectories were calculated at the individual level, from which a 0.84 potential predictability of locations could be inferred. All these findings arrive at a conclusion that residents' spatiotemporal behavior in working days is strongly regular. Furthermore, the article discussed the spatial heterogeneity of behavioral modes and predictability of locations, including characteristics of spatial distributions, as well as underlying influencing factors from the environment.