Urban flooding is a global challenge, that results in billions of dollars in recovery, taking thousands of lives annually around the world. Flooding has been shown to disproportionately affect marginalized populations, with those unable to prevent or combat the economic upheaval associated with repetitive flooding being impacted the most. Addressing the challenges of flood mitigation is not feasible without accurately delineating the extent of flooding; that is to define the communities impacted by specific floods. While recent advances in remote sensing and data availability open up new opportunities for automated analysis that is large in scale and fine in resolution, these opportunities have not yet been fully realized to delineate flood extent. Neighborhood residents and organizations contain a wealth of institutional knowledge regarding the extent and impacts of flooding that are not well integrated into existing flood delineation and impact assessments. Here we present a combination of both physical and social data to accurately account for flood impact on residents using two examples of flood prone communities from Philadelphia. Additionally, we present how flood risk and vulnerability shifts in cities across the US. This research moves towards developing methods for effectively and efficiently delineating flooding extents within the urban landscape at a resolution fine enough to identify those impacted by flooding, which is required to gain comprehensive solutions to urban flooding that currently remain out of reach.