Undergraduate Capstone Project - Bill Sun

Undergraduate Capstone Project

Vaccine and COVID-19 misinformation detection on social media, powered by natural language processing

Client: Public Health Ontario

September — December 2020


Working in a team of 4 and with a client, the goal of this capstone project is to help our client, Public Health Ontario, find a way to combat misinformation about vaccines and COVID-19 on social media, using a machine learning approach.

My team decided to use natural language processing techniques, especially sentiment analysis, to compute sentiment scores for each social media post. The data was obtained from Public Health Ontario, which consists of various labelled tweets on Twitter from earlier this year about COVID-19 and vaccination. We also decided to scrap our own data online using APIs from other social media platforms, including Facebook and Reddit.

We developed a web-based dashboard to display real-time statistics on how our sentiment analysis model rates the latest social media posts. These posts are scraped directly from the social media sites. A demo will be available upon completion.