Flooding in urban areas, especially in low-income or disadvantaged communities, poses a serious problem to drivers, even at depths usually associated with Nuisance Flooding. While techniques exist to map and predict flooding extent, a knowledge gap exists in accurate mapping and prediction of urban flooding. It is important that agencies and individuals be given an understanding of how much flooding a region may experience given a certain weather event, so that drivers may preemptively avoid flooded areas. This paper synthesizes several approaches to build an understanding of the spatial extent of urban flooding in the city of San Diego, California. First, flooding reported during major storms was used as validation data for a Generalized Linear Regression model to create a map of flood risk. Then, a Support Vector Machine model was used to extract areas of possible flooding from a COSMO-SkyMed image taken during a heavy storm. Finally, the performance of the original GLM model was compared with a new model that used both reported flooding and flooded areas extracted from the SAR image as inputs. Each model provided robust and meaningful results, the Generalized Linear Model indicating which areas of the city are most at risk for flooding and the image classification Support Vector Machine algorithm successfully picking out water bodies during both dry and wet conditions. A comparison between final results showed correlation between the two; areas found to be flooded during the storm were also areas found to be at a high risk for flooding.