Managing a natural disaster: actionable insights from microblog data

Abstract

Social media message boards have become a critical source of information during mass emergencies/disasters, leading to appropriate human action. The use of platforms like Twitter to share information about unfolding crises and social media adoption by governments for communication has increased interest in developing rounded disaster management strategies. Although scholarly works exist for modeling human-traits as social media usage predictors, seminal works on using social media as a predictor for human behavior are rare. This paper aims to identify pertinent information communicated amidst a disaster to unearth linguistic and thematic features that make tweets popular and attract human involvement. This research is based on the calamities during the last decade in the Indian subcontinent. We apply computational intelligence to identify features that make a tweet popular during a disaster. Our research suggests that Tweet popularity attracting human action in a disaster is affected by communication style over social media.