Cell phone data for quantifying disaster recovery

Keywords: resilience, access, hazard recovery, spatiotemporal data, mobile phone data
Abstract Type: Paper Abstract

Abstract

Natural and man-made hazards bring about changes in access to essential services such as grocery stores, health care, schools, and day care through closure of facilities, transportation system disruption, evacuation orders, power outages, and other barriers to access. Understanding access to essential services before, during, and after hazards is critical to ensure equitable recovery and more resilient communities. However, past approaches to understanding facility closures, inaccessibility, and household recovery such as surveys and interviews are labor-intensive and of limited geographic scope. Location-based-services (LBS) data from smartphones captures users’ behavior and mobility patterns, offering new opportunities for evaluating disaster response at large scales. Assessing LBS over the course of a hazardous event reveals deviations from behavioral patterns, capturing household- or facility-level response and recovery periods. I present an approach using LBS data and changepoint anomaly detection methods to identify changes in facility-level access that may indicate a closure or inhibited access to a facility. I demonstrate by analyzing loss of access to supermarkets, schools, health care facilities, and home improvement stores in Southwest Florida leading up to and following the landfall of Hurricane Irma in 2017. Next, I evaluate household level recovery by identifying home and workplace locations and typical visit patterns. I apply a novel Bayesian network-based anomaly detection method on these visit patterns to identify household-level recovery periods. I conclude with discussion of ongoing efforts to quantitatively assess the relationship between access to essential services facilities, household recovery, and community resilience.