Research Website
https://www.purdue.edu/discoverypark/vaccine/
Keywords
Visual analytics, social media, interactive, web, human behavior, disaster management
Presentation Type
Poster
Research Abstract
Real-time social media platforms enable quick information broadcasting and response during disasters and emergencies. Analyzing the massive amount of generated data to understand the human behavior requires data collection and acquisition, parsing, filtering, augmentation, processing, and representation. Visual analytics approaches allow decision makers to observe trends and abnormalities, correlate them with other variables and gain invaluable insight into these situations. In this paper, we propose a set of visual analytic tools for analyzing and understanding real-time social media data in times of crisis and emergency situations. First, we model the degree of risk of individuals’ movement based on evacuation zones and post-event damaged areas. Identified movement patterns are extracted using clustering algorithms and represented in a visual and interactive manner. We use Twitter data posted in New York City during Hurricane Sandy in 2012 to demonstrate the efficacy of our approach. Second, we extend the Social Media Analytics and Reporting Toolkit (SMART) to supporting the spatial clustering analysis and temporal visualization. Our work would help first responders enhance awareness and understand human behavior in times of emergency, improving future events’ times of response and the ability to predict the human reaction. Our findings prove that today’s high-resolution geo-located social media platforms can enable new types of human behavior analysis and comprehension, helping decision makers take advantage of social media.
Session Track
Data Trends and Analysis
Recommended Citation
Diego Rodríguez-Baquero, Jiawei Zhang, David S. Ebert, and Sorin A. Matei,
"Web-Based Interactive Social Media Visual Analytics"
(August 3, 2017).
The Summer Undergraduate Research Fellowship (SURF) Symposium.
Paper 87.
https://docs.lib.purdue.edu/surf/2017/presentations/87
Included in
Computational Engineering Commons, Computer Engineering Commons, Computer Sciences Commons, Electrical and Computer Engineering Commons
Web-Based Interactive Social Media Visual Analytics
Real-time social media platforms enable quick information broadcasting and response during disasters and emergencies. Analyzing the massive amount of generated data to understand the human behavior requires data collection and acquisition, parsing, filtering, augmentation, processing, and representation. Visual analytics approaches allow decision makers to observe trends and abnormalities, correlate them with other variables and gain invaluable insight into these situations. In this paper, we propose a set of visual analytic tools for analyzing and understanding real-time social media data in times of crisis and emergency situations. First, we model the degree of risk of individuals’ movement based on evacuation zones and post-event damaged areas. Identified movement patterns are extracted using clustering algorithms and represented in a visual and interactive manner. We use Twitter data posted in New York City during Hurricane Sandy in 2012 to demonstrate the efficacy of our approach. Second, we extend the Social Media Analytics and Reporting Toolkit (SMART) to supporting the spatial clustering analysis and temporal visualization. Our work would help first responders enhance awareness and understand human behavior in times of emergency, improving future events’ times of response and the ability to predict the human reaction. Our findings prove that today’s high-resolution geo-located social media platforms can enable new types of human behavior analysis and comprehension, helping decision makers take advantage of social media.
https://docs.lib.purdue.edu/surf/2017/presentations/87