Monitoring the Perception of Covid-19 Vaccine using Topic Models

Despite the unprecedented international collaboration to develop a Covid-19 vaccine at lightning speed, concerns remain whether the vaccine will be embraced by the population at large, calling into question its effectiveness at controlling the pandemic. This paper mines public outlook on the Covid-19 vaccine using unsupervised topic modeling. Tweets were collected a week following the announcement of Operation Warp Speed and were cleaned and labeled as anti-vaxx and pro-vaxx. Topic modeling was applied separately to these anti-vaxx and pro-vaxx groups of tweets. Anti-vaxx tweets lead to four themes: (i) lack of safety especially due to rushed development, (ii) conspiracy and conflict of interest, (iii) ideology, globalism and new world order, and (iv) loss of personal choice and freedom. Pro-vaxxers hope to convince skeptics by touting prior success of immunizations and express concern at the rise of anti-vaxx movement on Facebook and Twitter. Furthermore, they accuse anti-vaxxers of spreading falsehoods and misinformation. In addition to these concerns, however, pro-vaxxers mock, ridicule and insult anti-vaxxers in unflattering, disrespectful and derogatory tones. Public health implications of these themes are discussed in conclusion of the paper.