Text Mining of COVID-19 Discussions on Reddit

Syed Saad Naseem, Dhruv Kumar, Mohammad S. Parsa, Lukasz Golab

Abstract: We identify discussion communities on the Reddit social curation platform that frequently mention COVID-19related terms, and we apply the Non-negative Matrix Factorization (NMF) topic modelling algorithm to extract topics discussed by these communities. In addition to forums dedicated to COVID-19, we find such discussions on general question-answering forums, on a forum for teenagers, and on advice forums, specifically medical, mental health, relationship, and legal advice. Topic modelling results reveal the hardships of life during the pandemic and the effects of social distancing: fear of contracting the virus, job safety and security, loss of motivation and productivity, troubles with online activities, and dealing with those who do not follow social distancing rules.