Date of Award

Fall 2018

Document Type

Thesis

Degree Name

Master of Science in Information Systems (MSIS)

Department

Computer Science

College

College of Science

First Advisor

Sathish Kumar

Second Advisor

Zhenlong Li

Third Advisor

William Jones

Abstract

Disasters require quick response times, thought-out preparations, overall community, and government support. These efforts will ensure prevention of loss of life and reduce possible damages. The United States has been battered by multiple major hurricanes in the recent years and multiple avenues of disaster response efforts were being tested. Hurricane Irma can be recognized as the most popular hurricane in terms of social media attention. Irma made landfall in Florida as a Category 4 storm and preparation measures taken were intensive thus providing a good measure to evaluate in terms of efficacy. The effectiveness of the response methods utilized are evaluated using Twitter data that was collected from September 1st to September 16th, 2017. About 221,598 geotagged tweets were analyzed using sentiment analysis, text visualization, and exploratory analysis. The objective of this research is to establish an observable pattern regarding sentiment trends over the progression of the storm and produce a viable set of topic models for its totality. The study contributed to the literature by identifying which topics and keywords were most frequently used in tweets through sentiment analysis and topic modeling to determine what resources or concerns were most significant within a region during the hurricane Irma. The results from this study demonstrate that the sentiment analysis can measure people’s emotions during the natural disaster, which the authorities can use to limit the damage and effectively recover from the disaster. In this work, we have also reviewed the related works from the text/sentiment analysis, social media analysis from hurricanes/disaster perspective. This research can be further improved by incorporating sentiment analysis methods for classifying emoticons and non-textual components such as videos or images.

Share

COinS