Integration of Real Time Data in Urban Search and Rescue

In this pilot project, we demonstrate a preliminary workflow for incorporating real time, “crowdsourced” data into disaster relief routing, using search and rescue (SAR) operations in the aftermath of the 2010 Haiti earthquake as a test setting. We develop two simple models of relief routing: one with real time data and one without. We use road and building data from OpenStreetMap, a collaborative mapping project that assumed a prominent role in the postearthquake data dissemination process, to define road networks for routing analysis. The real time model incorporates daily updates to this road network contributed by volunteers around the world. For the real time model, demand is predicted by filtering text messages sent to the standardized shortcode 4636 for actionable information on collapsed structures and trapped persons. The non-real time model predicts demand using historical demand indicators such as building footprints.