New study uses social media data and ML to analyze COVID-19 response

A new research chapter from Nur Hafieza Ismail, Nur Shazwani Kamarudin, and Nurol Husna Che Rose (submitted to arXiv on June 9, 2026) examines how social media data can be systematically used to study the COVID-19 pandemic. With nearly 4.6 billion active social media users worldwide, these platforms generate massive amounts of real-time, unsolicited data. The authors categorize user-generated content into linguistic (language patterns), visual (images, memes), and emotional (sentiment, stress) indicators. They review deployed machine learning algorithms, natural language processing (NLP) techniques, feature engineering approaches, and survey methods that extract meaningful insights from this data. The study also outlines directions for future research, such as improving early outbreak detection and tailoring public health messaging.

The chapter emphasizes that social media, when used responsibly, can be a powerful tool for disseminating reliable news and raising public awareness among patients, clinicians, and society. By analyzing emotional and linguistic cues, researchers can better understand how different populations perceive and cope with the pandemic. The work is particularly timely as it demonstrates how unstructured social data can complement traditional epidemiological studies, offering a more granular view of public sentiment and behavior. While not peer-reviewed yet, this preprint provides a solid framework for leveraging social media analytics in public health crises, with implications for future infodemic management and crisis communication strategies.