Using Social Media to Improve Community Resiliency and Disaster Response
October 24, 2018
By Christine Wendling
For over 25 years, Margaret Reams has performed research focused on environmental policy and management, resilience, and sustainability. A professor of environmental sciences in LSU’s College of the Coast & Environment, or CC&E, Reams is bringing that expertise to the table with her latest research on the role social media plays in disaster communications and community resilience. In particular, she is looking at the difference in social media usage between Hurricanes Sandy (New York, 2012), Isaac (Louisiana, 2012), and Harvey (Texas, 2017).
Reams is part of a team of LSU and Texas A&M researchers, including Principal Investigator Nina Lam, professor of environmental sciences in CC&E; Michelle Meyer, assistant professor of sociology at Texas A&M; and LSU faculty from the Center for Computational Technology, or CCT, including Seungwon Yang and Seung-Jong Park. With funding from the National Science Foundation (NSF) they have created the Interdisciplinary Computation and Analysis of Resilience, or ICAR, project. As part of this project, they are analyzing nine million tweets from the time of the events to determine whether socio-economic and geographical disparities exist.
“We started with the assumption that there is a digital divide, it's sort of common sense, that the more urban areas are going to have more social media communication. And, Resilience Theory holds that more dense networks of communication among the stakeholders is a precondition of resilience,” Reams said.
These communication networks, among people who live and work together, enable them to function effectively. That, in turn, reduces their vulnerability by enhancing their ability to rely on each other and adapt to changing exposure risks quickly. After all, the very first people to respond to a disaster are those living within the local community.
“The question then becomes: is social media going to be used more where we see these dense networks of stakeholders or is there evidence that there are obstacles to its use? Like down here [in Louisiana] after Isaac. Because you were dealing with more rural areas, maybe some poverty issues, maybe it's an older population in the community, so maybe there would be natural barriers or challenges to really harnessing social media during these emergency phases. And, sure enough, we found that in 2012 it was no contest,” Reams said.
The LSU team found that, during Hurricane Sandy, New York had a rich network of social media activity that was not nearly as prevalent in Louisiana during Hurricane Isaac. Even NGOs and government agencies were not yet as social-media savvy as their northeastern counterparts.
“And, that was important to understand. So, we began looking at all kinds of differences between demographics to help explain variations in use of social media. We're still crunching through this data,” Reams said.
Once a month, Reams and Lam meet with the ICAR group who are building a machine-learning algorithm that will break down the millions of tweets into manageable categories, such as demographic distribution and the needs and sentiments being communicated by those demographics. Once they have that breakdown, this algorithm could become a powerful tool for understanding the sources, patterns, and consequences of social and geographical disparities. The researchers believe it will contribute important information about what steps communities can take to be better-prepared for future disasters.
Leveraging Social Media to Improve Disaster Management
In addition to categorizing the tweets, Reams and her PhD student, Ryan Kirby, are analyzing more than 300 responses to a survey they sent to NGOs and government agencies involved in emergency management during Hurricanes Sandy and Harvey. The survey results shed new light on how these organizations use social media for emergency management and the concerns they may have about social media as a tool for communication with the public. Reams and Kirby will present their findings at the American Geophysical Union 2018 Meeting in Washington, D.C. in December.
The good news is that, so far, they have found that governmental agencies and non-governmental agencies are organized and working symbiotically during the four phrases of emergency management: mitigation, preparedness, response, and recovery. Government agencies tend to be more active on social media during the first three phases of disaster management while NGOs are more active during the recovery phase. But, their use of social media is fraught with challenges. Many of the organizations say they are concerned about sharing inaccurate reports that would be more harmful than helpful during an emergency. Some have even implemented policies stating they will not share anything that does not come directly from Homeland Security. Another emerging problem they identified is how to respond to calls for help on social media.
“Some of the public agencies report receiving a tweet or Facebook post coming in saying, ‘I'm stuck on top of my house.’ That's really kind of a 911 call, and while they are legally required to respond to 911 calls, they don't have the capacity to follow up on all of the incoming social media posts or messages,” Reams said.
One potential solution would be to build self-organized groups within the community who can help. A well-known example is the Cajun Navy, an informal volunteer group of private boat owners, originally based in Louisiana, who assist in search and rescue efforts, most notably: the 2016 Louisiana floods, Hurricane Harvey in 2017, and Hurricane Florence in 2018. Reams believes anything the government can do to encourage self-organizing among the people will give them more “resilience capacity.”
Citizens are increasingly turning to social media for the latest information in the face of a disaster. This research may help to improve emergency management by establishing best practices for the effective application of social media during the different phases of an emergency, such as improving how disaster managers handle social media inquiries and calls for help, identifying the most appropriate timing for communications, and determining the best medium for them to get their message out.
Learn More About the Project
This is one component of a larger, on-going project led by Principal Investigator Nina Lam, Professor & Abraham Distinguished Professor of Louisiana Environmental Studies in the Department of Environmental Sciences, and funded by the National Science Foundation.
You can read more about the Twitter analysis in the recently published paper, “Mining Twitter Data for Improved Understanding of Disaster Resilience,” here: https://bit.ly/2QjCRr6
And, the full proposal and the rapid grant proposal for the project can be found here:
Full Proposal: https://bit.ly/2P0sr3f
Rapid Grant: https://bit.ly/2Pw15BD
Learn more about it in the recently published paper, “Mining Twitter Data for Improved Understanding of Disaster Resilience,” here: https://bit.ly/2QjCRr6