Understanding ordinary and disruptive events discussion in twitter: Barbados environmental health hazard as a use case

Abstract

Online users may utilize social media to discuss issues related to their environment. Understanding such discussions can help with predicting early warning signs for crisis situations and to enhance situational awareness and emergency preparedness. Over a period of more than four years, we collected and adopted a filtering method to obtain 30,358 tweets concerning environmental health risks in Barbados. In this study, we implemented an unsupervised machine learning algorithm to discover and understand how social media is used in the discussion of environmental health situational awareness. Moreover, what other topics online users are exposed to when engaging in such conversations. Our results show that there is a distinction between disruptive and ongoing events by exploring the number of tweets at a certain point of time and the sentiment of each event.