Download Presentation (2.1 MB)

Publication Date


Document Type


Presentation Type


Degree Type



Information Technology


Rosangela Follmann

Mentor Department

Information Technology


James Wolf

Co-Mentor Department

Information Technology


Social media has created an unprecedented way for individuals to share their concerns, fears, optimism, and happiness, for example, in ways that were not even conceivable some 20 years ago. Extensive data from these social media platforms, such as Twitter, makes it an invaluable resource for opinion mining and sentiment analysis. Starting in December 2019, the corona virus pandemic has had devastating consequences all over the planet, sparing no country. Health, social, and economic tolls associated with the pandemic has generated intense emotions and spread fear in people of all ages, genders, and races. During these difficult times, many have shared their feelings and opinions on many aspects of their lives via Twitter. In this project we use machine learning to measure subjectivity polarity in COVID-19 related tweets, labelling it as positive, negative, and neutral, depending upon the vocabulary encountered in the tweets. Our work focused on a detailed study of the distribution of opinions among the primary U.S. states. We also tested the relationship between the sentiment scores and the cases of COVID-19 in the United States, establishing a link between the sentiment scores, the reported cases and the death toll. The findings may assist with implementing legislation related to COVID-19, act as a reference for scientific work, inform and educate the public on critical pandemic-related issues.

Using Machine Learning To Measure Sentiment During The Covid-19 Pandemic