Analyzing Tweets to Understand Spatio-Temporal Variability of Situational Awareness Information

Authors: Bandana Kar*, Oak Ridge National Laboratory, Edwin Chow, Texas State University, Xiaohui Liu, University of Southern Mississippi
Topics: Geographic Information Science and Systems, Hazards and Vulnerability, Spatial Analysis & Modeling
Keywords: Data mining, disaster informatics, situational awareness, content analysis
Session Type: Paper
Day: 4/10/2018
Start / End Time: 10:00 AM / 11:40 AM
Room: Napoleon D1, Sheraton 3rd Floor
Presentation File: No File Uploaded

The ubiquity of social media has enabled the generation of large volume of data containing real-time information useful for emergency management activities. For instance, OpenStreetMap and Ushahidi were used during Haiti Earthquake (2010) to locate hot spots of crisis and to coordinate emergency response. However, these data tend to be noisy to be effectively used for disaster analytics. This study examined the following questions to determine the effectiveness of Twitter for disaster analytics: (i) what kind of information about situational awareness (SA) could be derived from tweets for emergency response? and (ii) what is the spatial and temporal distribution of such information throughout an emergency event? The preliminary results indicate that while a lot of non-relevant tweets were generated prior to September 12th (day of heavy precipitation during 2013 Colorado flood), tweets generated on and following September 12th contained crisis information and warnings. Limitations and future research direction is presented following these findings.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login