Computational analysis of Islamic values in digital children's programming: a BERTopic study of Quranic narratives on YouTube

The current paper is a computational model of analyzing the Islamic values and conflict representation in the digital programming used by children distributed over YouTube. Although some studies have expressed worries about what is being directed to the youth via digital media and the spread of Islamic education based media, the idea of systematic methodological analysis of religious content among Arabic children has not been fully examined. To analyze the animated series of stories based on the Quran (30 episodes, 15 stories of Quran), we use the topic modelling based on BERTopic, which is based on AraBERT and conflict classification. The value framework 6 Islamic values (Faith), Obedience, Justice, Mercy, Patience, Negative Values ) was developed by scientific discovery of data and has been proven by: (1) computational evidence (topic coherence = 0.67), and (2) scriptural outlier of the Quranic scripture. The classification system of conflict is of four types (Physical, Verbal, Moral, Divine Punishment), based on the narrative theory, inductive analysis, and research on media violence. BERTopic AarBERT is better than the old fashioned LDA (0.67 vs. 0.52 coherence) and the conflict classifier has an F1-score of 0.84. The analysis has identified that 85 percent of the conflicts involve determined moral resolution in which the content does not cater to the secular media due to absence of moral framing to the conflict. This framework fills the crucial gap between the development of Arabic NLP and the methodology of assessing the Islamic digital programming.