Understanding crisis communication on social media with CERC: topic model analysis of tweets about Hurricane Maria

Abstract

This study outlines the communication patterns that emerged throughout phases of Hurricane Maria according to the crisis and emergency risk communication model. Topic model analysis, latent semantic analysis, and word-cloud analysis were utilized to capture the complexity of communication through this extreme event. Topics that emerged during different phases of the event are discussed. Findings concerning communication use and content were consistent with previous research. However, the use of topic model analysis allowed for a more comprehensive understanding of communication during a natural disaster and uncovered communication patterns that are less represented in the existing literature. Implications and theory development in the case of crises alongside big data tools in crisis communication are discussed.