This project examines legal actions involving insurance companies that have denied claims following natural disasters, because geographical variation in litigation may help to explain why communities vary in their pace of recovery. The research examines demographic and economic factors that potentially underlie variation in litigation rates and the legal outcomes. The study also investigates correlations between litigation and recovery rates as indicated by construction that is evident over time in satellite images, which are processed with deep-learning algorithms. The findings and recommendations are made available to a broad range of stakeholders, including regional planners. The project also provides opportunities for the training and education of multiple students. As a contribution to the geographical sciences, this study examines the dynamic processes that lead to spatial variation in recovery rates after natural disasters. Communities may vary in their litigiousness, and this study uses qualitative methods to elucidate the reasons for considering legal recourse among a sample of impacted residents. Complementary methods are used to construct a quantitative parcel-level dataset in communities that have been negatively impacted by recent disasters. This dataset combines property records, court records, and satellite images of the study area. Deep learning algorithms are used with the satellite images to detect evidence of damaged homes and subsequent recovery. The da