UAH doctoral candidate leads study on leveraging AI and social media for disaster relief communication

In a groundbreaking study published in the International Journal of Production Research, a team led by Vishwa Vijay Kumar, a doctoral candidate at The University of Alabama in Huntsville (UAH), explores how social media platforms can be integrated with artificial intelligence (AI) to enhance communication between disaster victims and aid providers.

Kumar, pursuing his PhD in Industrial & Systems Engineering and Engineering Management at UAH, collaborated with Dr. Avimanyu Sahoo and Dr. Sampson Gholston, along with Dr. Siva K. Balasubramanian from the Illinois Institute of Technology.

The research focuses on analyzing data from X, formerly known as Twitter, during two critical periods of the COVID-19 pandemic: the initial outbreak in the United States (March-April 2020) and the delta variant surge in India (May-June 2021). These times were marked by significant disruptions in healthcare supply chains, leading to shortages of vital medical supplies such as face masks, medicines, and ventilators.

Kumar’s motivation for the study stems from his personal experiences growing up in Sitamarhi, Bihar, a flood-prone region in India.

“Natural disasters frequently isolate communities, making it crucial to develop a communication framework that connects those in need with the outside world and authorities,” Kumar explained. The pandemic underscored similar challenges on a global scale, highlighting the need for efficient resource allocation and real-time information sharing.

The research team devised a four-step process utilizing AI and machine learning to analyze 3.9 million tweets.

They identified keywords to classify tweets as either “imperative” (actionable requests for help) or “non-imperative” (non-actionable information).

The study also included algorithms to estimate the geographic locations of imperative tweets lacking geo-tags, facilitating more effective coordination of aid.

Future research will address various challenges identified during the study, such as improving geo-location accuracy for social media posts and forecasting supplies of vaccines, healthcare, and food.

The team also plans to explore the use of other social media platforms like Facebook and Instagram, and the applicability of their methods to other disaster scenarios, including hurricanes and earthquakes.

Kumar and his colleagues aim to develop a real-time platform to scan social media posts during disasters, generating immediate reports on supply and demand issues and pinpointing the locations of people requesting help.

This innovative approach promises to significantly improve disaster response and resource management in future crises.

For more information, please visit www.uah.edu/news.