UNVEILING SENTIMENTS AND EMOTIONS IN THE DIGITAL DISCOURSE: ANALYZING YOUTUBE COMMENTS ON THE WESTERN SAHARA CONFLICT

Abstract

This research investigates the emotional dynamics within YouTube comments related to the Western Sahara conflict, with a focus on sentiments, expressions of fear, and a broad spectrum of emotions. The study analyzes a dataset of 10,000 Arabic comments from the top 10 most viewed YouTube videos concerning the conflict, collected over the last three years, employing both human and ChatGPT-generated annotations. The findings reveal a complex emotional landscape in which pessimism predominates, reflecting the enduring challenges of the conflict, while hopeful sentiments persist, signaling optimism for peaceful resolutions. Remarkably, expressions of fear were relatively low, indicating that fear is not the predominant emotion in these discussions. The diversity of emotional expressions, encompassing anger, disgust, enjoyment, sadness, and surprise, underscores the multifaceted nature of how individuals engage with the Western Sahara conflict. However, the alignment between human and ChatGPT annotations varies across sentiment categories and emotional expressions, highlighting the challenges and potential of automated sentiment analysis in complex geopolitical discourse. This research contributes to our understanding of the emotional dimensions of online geopolitical discussions and calls for further exploration into the alignment of online sentiments with offline public opinion.

Date of Award	2024
Original language	American English
Awarding Institution	HBKU College of Humanities and Social Science