Through microblogging applications, such as Twitter, people actively document their lives even in times of natural disasters such as hurricanes and earthquakes. While first responders and crisis-teams are able to help people who call 911, or arrive at a designated shelter, there are vast amounts of information being exchanged online via Twitter that provide real-time, location-based alerts that are going unnoticed. To effectively use this information, the Tweets must be verified for authenticity and categorized to ensure that the proper authorities can be alerted. In this paper, we create a Crisis Message Corpus from geotagged Tweets occurring during 7 hurricanes in the United States. Using this annotated corpus, we train a machine learning classifier to identify requests for help in real time. Through a deep learning model, we remove tweets that are below our classification confidence threshold of 98%. Using this model in conjunction with a front-end dashboard can allow service teams in crisis areas to be notified of alert-tweets without having to sift through hundreds of non-relevant tweets.

Creative Commons License

Creative Commons Attribution-Noncommercial 4.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial 4.0 License