The Virulence of a Tweet: The Effects of Twitter Language on COVID-19 Spread

Abstract

The rise of COVID-19 throughout the world has created unprecedented circumstances and new, rich datasets containing information to understand the outcomes of the challenges. In an attempt to understand how certain states in the United States handle COVID-19 more effectively than others, sentiment analysis looking at Twitter data, policy data and COVID-19 count data was conducted. By looking at the sentiment of tweets along with case numbers per State the goal of this project is to identify if there is a correlation between sentiment of COVID-19, specifically gauged through opinions on masks, and how the state is doing in total cases. With this analysis, data was gathered to identify which language that was helpful towards stopping the spread of COVID-19 and promoting the usage of masks. We are able to find events that are correlated with an increase in positive sentiment about mask wearing as well as the most frequent words used to describe mask wearing. This will give policy and law makers insight into what language was used to create effective policies during unprecedented times and perhaps a new approach as to how they should be addressing global challenges by using similar language or words. With the development of the vaccine we will include looking at the language surrounding it to see the varying opinions about getting the vaccine and how politicians address the public’s qualms.