Identifying Hurricane Evacuation Intent on Twitter

Evacuations have a significant impact on saving human lives
during hurricanes. However, as a complex dynamic process, it
is typically difficult to know individual evacuation decisions
in real-time. Since a large amount of information is continuously posted through social media platforms, we can use them
to understand individual evacuation behavior. In this paper,
we collect tweets during Hurricane Irma in 2017 and train a
text classifier in an active learning way to distinguish tweets
expressing positive evacuation decisions from both negative
and irrelevant ones. Additionally, we perform a demographic
analysis and content clustering to investigate the potential
causes and correlates of evacuation decisions. The results can
be used to help inform planning strategies of emergency response agencies.