Mining Twitter Data to Understand the Human Sentiment on Hurricane Florence

Abstract

Introduction: Most studies have analyzed how natural disasters exert a severe impact on the regional level in the disaster period based on quantitative methods. This study aimed to highlight how Hurricane Florence exerts an impact on human life and societies across US states in a multitude of periods by employing both qualitative and quantitative methods.

Method: This study developed a new app called “Twitgis,” collected 1,433,032 tweets, and employed 57,842 data filtered for Hurricane Florence between 08-21-2018 and 10-01-2018.

Results: First, this study showed that the spatial patterns of tweets are differentiated by periods. For example, the spatial patterns of tweets are more concentrated in the south region in the pre-hurricane period, the spatial patterns of tweets are heavily concentrated in the Southeast region in the hurricane period, and the spatial patterns of tweets are more located in the Northeast region in the post-hurricane period. Second, the most retweeted tweet shows that human sentiment plays an important role in disaster information more than news of the hurricane in online communication. The first ranked tweet is about two times higher than the sum of the retweet numbers between the top two and top 20. Third, this study found that people actively utilize Twitter to share a lot of emotions, opinions, information, and so on for Hurricane Florence. For instance, about one-fifth of tweets in the sentiment analysis are emotions for the hurricane event.

Conclusion: Governments and policymakers should monitor Twitter data to understand the effects of natural disasters on people and the human environment.