Movement-based disruption estimators: Using mobile location data to predict community variation in disaster impacts

Abstract

Predicting which communities will be most disrupted by natural or anthropogenic disasters is of central concern to strategic planners seeking to optimize equitable outcomes of infrastructure investment. In this paper, we describe an approach to using mobile location data to estimate the relative magnitude of disruption across communities with arbitrary boundary delineations and use predictive modeling to show how mobility metrics and Census-based demographic information can be combined to predict the impact of similar disasters in novel scenarios. We demonstrate our approach through application of the proposed methodology to the Colonial Pipeline hack of 2021 and discuss opportunities for alternatives and refinements given additional data sets. The resulting movement-based estimation and prediction approach offers an avenue for ensuring a more resilient nation through strategic planning.