Investigating the agenda of global warming on Twitter: A machine learning approach

This study investigates the agenda of global warming issue based on the two levels of agenda-setting framework. Using the Bidirectional Encoder Representations from Transformers (BERT)-based models, it identifies nine attributes of global warming issue and estimates the average tone of tweets in each month of 2021. The findings suggest the attributes are interconnected and partisan divides existed to a larger extent in certain attributes than in others. The nine attributes identified, their inherent interconnections as well as different tones associated with these attributes have the implications for policy-makers to formulate and evaluate policies regarding this multifaceted issue.