Category: Data Science
MSDS Student on Winning Team for GSG Hackathon
May 18, 2021 D Sinopoli and P Kosi with RJ Braun | Photo by
Dominick Sinopoli, MSDS 2021 Graduate.
On the weekend of April 23, 2021, UD’s Graduate Student Government (GSG) sponsored an online hackathon. The overall theme of the hackathon was misinformation and cybersecurity. From the hackathon website[https://sites.udel.edu/gsg/gsg-hackathon], “the goals of the Hackathon are to provide the participants with a platform to: pitch their ideas; learn new skills and improve presentation skills; network with students from various backgrounds; have fun developing and coding new tools; improve their CVs; win prizes and build industry connections…” The subject areas included public health, finance, media, and cybersecurity. There were ten projects in the hackathon, taken on by graduate students representing a broad range of majors and colleges across campus.
MSDS student Dominick Sinopoli and Mechanical Engineering graduate student Prosper Kosi Anyidoho teamed up to work on the propagation of misinformation about COVID-19 in social and conventional news media online.
Their project tried to tackle the problem of misinformation during COVID-19. Due to social media, news outlets, etc., many statements and much information pertaining to COVID is spread easily, but how reliable and accurate is this information? The goal of the project was to try and develop a machine learning-based fact-checker. To achieve this goal a simple natural language processing model was built to distinguish facts from misinformation. Labeled data was obtained from the CodaLab Covid19 fake news detector competition. It included scraped tweets from Twitter and other social media platforms. Once the model was trained it was plugged into a simple webpage built with Flask that would allow users to type in their own statements. Through this process, the webpage became a COVID fact-checker. A user could input a statement, then a label of True or False would be returned. The result displayed is based on the model prediction for the entered statement. With some additional work, the fact-checker webpage could be very beneficial to people when they read an article or listen to the news and want to know if a statement is true or false.
Mechanical Engineering graduate student Prosper Kosi Anyidoho.
Congratulations to Dominick and Prosper on producing the winning project!