Sentiment Analysis on COVID-19 Related Tweets
Arianna Cao, Claire Zheng, Harshini Magesh, Dominic Crisafulli
Abstract
The recent COVID-19 pandemic has drastically altered life on a global scale. Authorities rely on accurate reports of confirmed cases to make decisions on the well-being of citizens. During these times, there is also a growing reliance on social media for citizens in quarantine to communicate their opinions and sentiments on their current situations. This project applies models such as Logistic Regression, Support-Vector Machine, Naive Bayes, and Stochastic Gradient Descent to classify COVID-19 related tweets with a sentiment of -1, 0, 1, meaning negative, neutral, and positive, respectively. In doing so, we found an inverse relation between the average sentiments of Twitter users in a specific location to the number of cases in that area at a given date.
Github Link
Our CodeOur Poster
Word Bubbles
Positive Sentiment
Negative Sentiment
.png)
Neutral Sentiment
Our Infographic
