Clustering reveals key behaviours driving human movement network structure

Abstract

AbstractWe analysed daily travel patterns from in-app GPS data in the United Kingdom to identify characteristic modes of travel behaviour, and the relative importance of different behaviours for the topology of the overall travel network. We clustered the detailed travel trajectories and identified four characteristic modes of travel that we named:regular-travel, andstay-at-homewhich represented the majority of travel days, and two ‘long tailed’ travel modes:long-distance, andaway-from-home, which represented ∼2.6% of travel days. We focused on these second two modes, in which individuals travel a long distance from home and deviate from predictable routines. We demonstrated how this relatively small portion of travel behaviour plays an outsized role in influencing the connectivity of the aggregate travel network. This analysis highlights the need to understand individual behavioural variations that can dictate the collective dynamics recorded by aggregate descriptions of human travel.