Crisis Communication of Public Agencies in Twitter: A Case Study of Hurricane Irma Exploring the Relationship between Topics and Engagement

Abstract

This study analyzes the crisis communication posts of public agencies on Twitter during a major natural disaster, Hurricane Irma. Analyzing engagement metrics of tweets’ relationship with factors such as communication topics, tweet and Twitter account characteristics, and time of posting tweets, the research aims to uncover effective aspects for enhancing public engagement and response during extreme events. Using machine learning and statistical approaches, we analyze tweets from prominent public agencies active in response to Hurricane Irma. An engagement metric, such as the number of retweets, was used to measure the effectiveness of crisis communication posts. Results indicate that real-time updates on storm prediction, preparedness activity, user concern and recovery gained higher engagement. Besides posting frequent tweets by agencies, time of tweeting and an agency’s popularity measured by the number of likes and followers, level of the agencies (e.g., federal, or regional) are associated with higher engagement. The study emphasizes the significance of efficient messaging and clear communication in capturing public attention during crises. It provides valuable insights for public and emergency management agencies seeking to improve their crisis-related social media strategies, specifically on Twitter. This study also assists public agencies in refining their social media communication strategies for future crises by identifying key elements of successful engagement in social media.