Modeling aggregate human mobility patterns in cities based on the spatial distribution of local infrastructure

Abstract

Understanding human mobility patterns in urban areas is key to solving a wide range of socio-technical problems at the human-infrastructure interface. Extending the intervening opportunities concept, we showcase a data-driven, network-based model that reproduces aggregate mobility patterns in cities. Using this model, we create a digital replication of daily travel across different trip purposes in 5 U.S. metropolitan areas and compare results against publicly available reference data. We find that our proposed model explains a large fraction of the variation in mean and median travel distance across the 5 cities. In particular, it accurately captures the effect of density on aggregate travel patterns. These findings add to evidence that human mobility patterns are strongly governed by the structure of the built environment. We discuss implications for the ongoing transformation of cities and for developing more sophisticated models that replicate human behavior based on crowd-sourced, spatio-temporal data streams.