Authors: Guimin Zhu*, Center for Geospatial Infornation Science, Department of Geographical Science, University of Maryland - College Park, Kathleen Stewart, Center for Geospatial Information Science, Department of Geographical Science, University of Maryland - College Park, Junchuan Fan, Oak Ridge National Laboratory
Topics: Geographic Information Science and Systems, Medical and Health Geography
Keywords: mobility change, COVID-19, Random Forest, spatial modeling
Session Type: Virtual Paper
Start / End Time: 8:00 AM / 9:15 AM
Room: Virtual 8
Presentation File: No File Uploaded
The COVID-19 global pandemic has highly impacted the U.S., with about 7.8 million positive cases and over 210,000 deaths occurred since January 2020. Additional peaks in both cases and deaths are predicted by health experts for this coming fall and winter. Florida has experienced a high incidence of COVID-19, accounting for approximately 10% of the total COVID-19 cases in the U.S. We have been investigating the relationships between human mobility changes and COVID-19 positive cases at census tract level for two counties in Florida, Miami-Dade and Broward Counties. A Random Forest regression model has been constructed using a set of mobility- and socioeconomic-related variables as well as COVID-19 positive case data to investigate the impact of these factors on mobility changes (i.e., the change in total number of inflow trips during COVID-19). While mobility in these two counties continued at almost-normal levels from late March to June 2020, as cases trended steeply upwards in June, mobility did show signs of change, especially in census tracts where cases were highest. The Random Forest model showed that factors including, e.g., distance to beach, level of education, age, as well as race and ethnicity, were returned as key variables associated with the changes to mobility. The significant explanatory variables will be helpful for local public officials tasked with developing mitigation strategies to impede the spread of COVID-19.