Modeling wildfire mobility response: Using a data-driven approach to optimal resource allocation

This project examines the effects of natural disasters on the mobility of individuals. Owing to impacts on critical infrastructure and local economies, people often change how and where they move following natural disasters. Understanding these changes is important for emergency response and long-term urban resilience planning. In this project, the researchers develop advanced computational solutions based on spatial analytics and locational data, such as information from mobile phones, to study how communities respond to disasters, especially wildfires. By looking at anonymous movement patterns, the project explores how people react to emergencies and how their ability to move varies across a geographical region. Open-source tools are developed that make it easier for researchers to investigate disaster impacts. These tools help to improve decision-making in urban planning and emergency response. In addition, the project provides students with opportunities to develop valuable analytical skills, helping to prepare them for future careers in STEM. There is a need to understand better how disasters affect people’s movements in real time, especially at a local level during different stages of an event. To address this need, the project creates an integrated spatial analytical and modeling framework that combines large-scale anonymous movement data, such as data from mobile phones, with geographic, demographic, and environmental information in order to study travel patterns and logistical vulnerabilities during crisis events. Specifically, the project pursues three goals. First, it uses different types of mobile phone data as behavioral markers to learn and model how people’s movement patterns change during various phases of a disaster, with particular attention on access and placement of resources. Second, the project develops new machine learning and spatial analysis methods to enhance the representation of mobile location data for modeling and summarizing population-level movement patterns. Third, the researchers are implementing and evaluating open-source software that integrates the methods for use by practitioners. The case studies assess how well the new methods can elucidate geographical areas where people may have trouble accessing needed resources during natural disasters. The results help to improve predictions of how people behave in emergencies and support better assessments of community responses based on data-driven insights. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.