Social media-based demographic and sentiment analysis for disaster responses.

Abstract

This study explores disaster responses across the United States for Winter Storm Jaxon in 2018 by utilizing demographic and sentiment analysis for Twitter®. This study finds that people show highly fluctuated responses across the study periods and highest natural sentiment, followed by positive sentiment and negative sentiment. Also, some sociodemographic and Twitter variables, such as gender and long text, are strongly related to human sentiment, whereas other sociodemographic and Twitter variables, such as age and the higher number of retweets, are not associated with it. The results show that governments and disaster experts should consider a multitude of sociodemographic and Twitter variables to understand human responses and sentiment during natural disaster events.