probabalistic

daiwikpal/Transmission_Risk

Currently, there are 26.2 million COVID-19 cases in the US due to people taking lower precautions to reduce transmission at public venues. This project aims to create a tool that provides individuals with information to make a responsible decision about visiting an establishment to prevent unnecessary SARS-CoV-2 cases. A probabilistic model that predicts the risk of an individual contracting COVID-19 (transmission risk) during a visit to an establishment was created. This model was integrated into a web app built using the StreamLit framework in Python3.8. It acquired data through parsing COVID-19 databases and obtaining venue-specific data through user-inputs. The transmission risk model worked by multiplying the likelihood an individual will encounter a virus carrier (exposure risk) and the risk the individual will contract the virus if one carrier is present (contract risk). The exposure risk model was tested through the Spearman’s Rank Correlation by calculating the correlation between the model’s risk prediction for 200 random counties to new cases two weeks later of the aforementioned counties. Contract risk was tested using four different scenarios. Transmission risk was tested using these four scenarios factoring three counties with ranging incidence rates. The exposure risk model averaged a Spearman Rank correlation of 0.81, placing it in the “very strong” category. The contract and transmission risk provided sensible predictions for the scenarios provided. This model can be easily expanded to other databases and adapted to high-incidence countries. Since virus-specific aspects can apply to other illnesses, the model can be adjusted easily for other viruses.