Climate change has become a significant threat to our society. Researchers in climate change communication often rely on specific program-based interventions to train public communicators, such as journalists, to raise awareness about critical issues and events. However, researchers face challenges in analyzing the change in communication behavior of participants in the programs over time since using traditional social science research methods alone, such as surveys and interviews, for analyses could be resource-intensive and costly. Social media mining provides an opportunity for such communication researchers. In this paper, we propose a social media mining system called Evo-Lyzer to measure and analyze evolving communication behavior on social media. Evo-Lyzer employs topic modeling and novel measures of communication divergence and consistency, which are informed by prior research of communication sciences, to evaluate an intervention program for climate change communication. We showcase Evo-Lyzer's capability in analyzing real-world data and assessing the effectiveness of a climate change communication program aimed at journalists and weathercasters. The application of our social media mining system could empower communication researchers to conduct large-scale studies on evolving communication behavior in program-specific interventions across different domains for sustainability and social good beyond communication about climate change.

PDF