Dimensions of the Use of Volunteered Geographic Information in Mass Crisis Events

Abstract

Recent studies have suggested that catastrophic events that trigger mass evacuation require surrounding communities to be well-prepared to act as ingress or pass-through areas for potential evacuees; however surrounding rural communities may have insufficient disaster-related logistical resources. In the response phase of disaster management, officials must be able to deploy resources to demand locations in types and quantities based on real-time requirements. Effective cross-jurisdictional disaster management needs real-time information, which is usually unavailable from official, authoritative sources. Conversely, VGI (volunteered geographic information) has the capability to provide real-time and local information in disaster management. This study investigates the possibility of utilizing real-time or near real-time VGI in mass evacuation scenarios. The study identifies a potential VGI data source, Tweets from Twitter and how to search for, discover and select relevant Tweets. The dissertation proposes research methods for harvesting, managing live Tweets and saving them to a distributed geodatabase for further spatio-temporal analysis and dissemination to users, such as responders and evacuees.;The study implements a Web GIS application, which includes a tweets discovery component, a geo-tagged tweets mapping component, and an online geo-tagged tweets operation component. The major research goals include designing an application programing interface (API) to harvest relevant Tweets and implement a distributed geodatabase system for storage, analysis, and display of the harvested Tweets so that vital information can be distributed in near real-time. Two case studies, based on Super Storm Sandy in 2012 and a shooting at Kent State University in 2014, were used to evaluate the pros and cons of Tweets from Twitter for response in emergency management and offered prototypes for the development of the final on-line Web GIS.