by Paul Govern
Reported in the Journal of Medical Internet Research, a statistical analysis of sentiments expressed online by U.S. Twitter users captures the rural-urban divide regarding COVID-19.
Google software engineer Yongtai Liu, PhD’22 in computer science; Bradley Malin, professor of biomedical informatics and computer science, and colleagues created a natural language processing system backed by artificial intelligence to identify changing patterns of sentiment concerning COVID in 407 million tweets containing geolocation data, posted from May 2020 to January 2022.
Negative sentiments about COVID prevention and vaccination ran stronger in rural areas. Dismissal of misinformation and conspiracy theories was stronger in urban areas. Rural users were more apt to criticize White House medical advisor Anthony Fauci and decry the pandemic as planned or fraudulent.
The authors say the findings demonstrate the utility of social media for monitoring public sentiment and targeting pandemic prevention and management.
Currently, the Centers for Disease Control and Prevention reports a cumulative COVID death rate in nonmetropolitan areas 34% higher than that in metropolitan areas (433 and 322 per 100,000, respectively).
Others on the study include Zhijun Yin, assistant professor of computer science; Congning Ni, computer science graduate student; and Chao Yan and Zhiyu Wan, postdoctoral research fellows in biomedical informatics.
The study was supported in part by the National Institutes of Health (HG009034).