Mobile homes are an affordable housing option for many Americans, however the infrastructure of these homes make these communities more susceptible to harm from tornados. In this research, geographic information science (GIScience) and geospatial artificial intelligence (GeoAI) analysis were used to determine if deep learning algorithms can recognize mobile home parks from aerial imagery covering six counties in Georgia. A multi criteria decision analysis (MCDA) was also performed to assess the social impacts of tornados on vulnerable populations in Georgia. Data layers of mobile home parks, proximity to previous tornado tracks and socio-economic characteristics including race and income from the U.S. Census were combined to examine the disproportional social impact of tornadoes on residents of mobile home parks. The methodology developed in this research will allow local planners and disaster responders to identify and recognize local areas with vulnerable populations requiring disaster preparedness, response and recovery due to damaging tornados.