The COVID-19 pandemic impacted just about every aspect of our daily life. It even ravaged our mood online.
That’s the result of a study from researchers at IU and the Luddy School of Informatics, Computing, and Engineering, who analyzed Twitter timelines for users in 20 metropolitan areas from January-July 2020 and found a linear relationship between worsening COVID-19 city data and online sentiment. Their paper, “Declining well-being during the COVID-19 pandemic reveals U.S. social inequities,” which was published in PLOS ONE, also provides insight into how the pandemic disproportionally affected communities of color and other at-risk groups.
“Our group continuously monitors societal well-being and public health related factors from large-scale Twitter data, sometimes in real-time,” said Johan Bollen, a professor at the Luddy School who worked with colleagues from the IU School of Public Health, and the Department of Psychological and Brain Sciences on the study. “Last year, we saw a tremendous, unprecedented drop in societal well-being at the time of the murder of George Floyd. This led us to explore how societal well-being changed during the COVID-19 pandemic.”
The group compared changes in Twitter sentiment over time across the cities and correlated declines in well-being with COVID-19 infection rates, population, population density, and city demographics, showing that COVID-19 infections led to significant drops in well-being. Researchers discovered the relationship was strongly affected by the demographics of the respective cities, and, in fact, residing in predominantly white cities had a strong and persistent protective effect against COVID-19 driven negative mood.
Real-time quantification and tracking of how socio-economic changes impact communities can help shape public policy and lead to more equitable outcomes.
“We can’t have a one-size-fits-all approach to public health because where you live and the demographics of that area have a significant effect on how you are affected by a public health crisis, even if it’s the exact same virus and disease,” Bollen said.
The researchers plan to develop more fine-grained, real-time systems to track not just well-being but also a variety of indicators of health and mental health across the full spectrum of available indicators.
“The insights that can be gleaned from social media data are expanding the tools that can be used to improve lives,” said Kay Connelly, the associate dean for research for Luddy. “Professor Bollen and his group are the perfect example of how innovative analysis of technology and the data at our disposal have the potential to make a real-world impact.”