Crisis Media - Discovering Social Media's Role in the Emergency Management Process
Open Access
- Author:
- Doty, Christopher
- Graduate Program:
- Informatics (PHD)
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- January 30, 2024
- Committee Members:
- Jeffrey Bardzell, Program Head/Chair
Prasenjit Mitra, Major Field Member
Timothy E. Kohler, Special Member
Stephanie Madden, Outside Unit & Field Member
Andrea Tapia, Major Field Member
Nick Giacobe, Chair & Dissertation Advisor - Keywords:
- Social Media
Data Mining
Interviews
Emergency Management
Crisis
Disaster - Abstract:
- The use of social media platforms has opened up a new avenue for information gathering by those in the emergency management field. This dissertation examines the use of this new source of information and seeks to add to the Crisis Informatics field of research. To do so, semi-structured interviews were conducted with members of the Emergency Management community. These individuals play a vital role in the disaster response and recovery process. They have the responsibility of providing information, gathering resources, and coordinating response efforts when disaster strikes. The aim of these interviews was to discover what information these emergency managers believed could be obtained from social media in order to aid in their operations. The coding of the interview transcripts found a set of “information categories” that appeared across the responses of the participants that made up the sample. These categories have been used to describe the types of information the emergency managers indicated as being useful. From there, the categories are used to justify a list of design considerations that future researchers should consider when designing social media analysis tools. Alongside the creation of design considerations was the development of different prototyped social media data analysis tools. This development was completed to attempt to conduct real time analysis of social media data. These tools utilized the Natural Language Processing methods of sentiment analysis, named entity recognition, and term frequency analysis. When applied to various datasets of collected Tweets, a potential example of event detection was created.