Measuring COVID-19 Related Media Consumption on Twitter
share
The COVID-19 pandemic has been affecting the world dramatically ever since 2020. The minimum availability of physical interactions during the lockdown has caused more and more people to turn to online activities on social media platforms. These platforms have provided essential updates regarding the pandemic, serving as bridges for communications. Research on studying these communications on different platforms emerges during the meantime. Prior studies focus on areas such as topic modeling, sentiment analysis and prediction tasks such as predicting COVID-19 positive cases, misinformation spread, etc. However, online communications with media outlets remain unexplored on an international scale. We have little knowledge about the patterns of the media consumption geographically and their association with offline political preference. We believe addressing these questions could help governments and researchers better understand human behaviors during the pandemic. In this thesis, we specifically investigate the online consumption of media outlets on Twitter through a set of quantitative analyses. We make use of several public media outlet datasets to extract media consumption from tweets collected based on COVID-19 keyword matching. We make use of a metric "interaction" to quantify media consumption through weighted Twitter activities. We further construct a matrix based on it which could be directly used to measure user-media consumption in different granularities. We then conduct analyses on the United States level and global level. To the best of our knowledge, this thesis presents the first-of-its-kind study on media consumption on COVID-19 across countries, it sheds light on understanding how people consume media outlets during the pandemic and provides potential insights for peer researchers.
READ FULL TEXT