Vishwa Vijay Kumar, a Ph.D. candidate at The University of Alabama in Huntsville (UAH), led a study published in the International Journal of Production Research, exploring how social media platforms can be utilized with artificial intelligence (AI) to enhance communication and support for disaster victims. This study, conducted with co-researchers Dr. Avimanyu Sahoo and Dr. Sampson Gholston from UAH, and Dr. Siva K. Balasubramanian from the Illinois Institute of Technology, focuses on bridging the gap between disaster victims and outside aid.
Research Context and Motivation
Kumar's inspiration stems from his upbringing in Sitamarhi, Bihar, India, a region frequently afflicted by floods. The recurring natural disasters in his hometown sparked his interest in creating a framework to facilitate communication between those in need and potential rescuers. The COVID-19 pandemic, which severely disrupted global healthcare supply chains, reinforced Kumar’s resolve to explore social media and AI as tools for faster disaster response and effective resource allocation.
Study Methodology
The research analyzed data from Twitter (now known as X) during two critical periods: March–April 2020 in the United States and May–June 2021 in India. These timeframes corresponded to the initial outbreak of COVID-19 and the surge of the delta variant, respectively. The disruptions in healthcare supply chains during these periods resulted in acute shortages of essential medical supplies.
Data Analysis and Findings
Using AI and machine learning algorithms, the researchers processed 3.9 million tweets, identifying keywords to distinguish between actionable pleas for help ("imperative" tweets) and non-actionable information ("non-imperative" tweets). The study also estimated the geographic locations of imperative tweets that lacked geo-tag information to enhance coordination of aid efforts.
The analysis highlighted several challenges in healthcare supply chains during disasters, which are proposed as areas for future research:
1. Geo-location of Non-geo-tagged Posts: Identifying the locations of individuals seeking help who did not provide their geographic information.
2. Supply Forecasting: Forecasting the availability of critical supplies such as COVID-19 vaccines and other health and food resources.
3. Use of Multiple Social Media Platforms: Extending the research to include other platforms like Facebook and Instagram to improve disaster response strategies.
4. Application to Various Disasters: Applying the developed techniques to other disaster scenarios such as hurricanes and earthquakes.
Future Developments
The research team plans to create a platform that will scan social media posts during disaster events to generate real-time reports on demand and supply issues, along with the geo-locations of individuals requesting assistance. This tool aims to enhance the efficiency and speed of disaster response operations.
The study underscores the potential of integrating social media and AI to address supply chain challenges during disasters. By leveraging these technologies, the researchers aim to improve the communication and coordination of aid, ultimately mitigating the impact of disasters on affected populations.
More: https://techxplore.com/news/2024-07-social-media-ai-chain-disasters.html
