Abstract
This study highlights spatial responses and displacements following Hurricane Michael by analyzing X® (formerly known as Twitter) data across various states in the United States over a multitude of periods, regions, and demographic characteristics. This study finds that the top three states damaged by the hurricane (Florida, Georgia, and North Carolina) show a significantly higher proportion of tweets during the hurricane week than the prehurricane and post-hurricane weeks. Furthermore, people's responses on social network services have some similar and dissimilar characteristics compared to the census data. For instance, Hispanics show significantly low responses to the hurricane event in all the three states compared to other races/ethnicities, as reflected in the census data. Finally, the gender and region variables play an important role in displacements within the binary logistic regression model, whereas the race/ethnicity and age variables are not related to displacements. To be specific, females show 0.4 times more displacements than males, and Georgia and North Carolina exhibit 1.3 and 0.8 times more displacements than Florida.