Since its inception, YouTube has been a source of entertainment and education. Everyday millions of videos are uploaded to this platform. Researchers have been using YouTube as a source of information in their research. However, there is a lack of bibliometric reports on research carried out on this platform and the pattern in the published works. This study aims at providing a bibliometric analysis on YouTube as a source of information to fill this gap. Specifically, this paper analyzes 1781 articles collected from the Scopus database spanning fifteen years. The analysis revealed that 2006-2007 were initial stage in YouTube research followed by 2008 -2017 which is the decade of rapid growth in YouTube research. The 2017 -2021 is considered the stage of consolidation and stabilization of this research topic. We also discovered that most relevant papers were published in small number of journals such as New Media and Society, Convergence, Journal of Medical Internet Research, Computers in Human Behaviour and the Physics Teacher, which proves the Bradford's law. USA, Turkey, and UK are the countries with the highest number of publications. We also present network analysis between countries, sources, and authors. Analyzing the keywords resulted in finding the trend in research such as "video sharing" (2010-2018), "web -based learning" (2012-2014), and "COVID -19" (2020 onward). Finally, we used Multiple Correspondence Analysis (MCA) to find the conceptual clusters of research on YouTube. The first cluster is related to user -generated content. The second cluster is about health and medical issues, and the final cluster is on the topic of information quality.
The COVID-19 pandemic introduced unprecedented challenges for people and governments. Vaccines are an available solution to this pandemic. Recipients of the vaccines are of different ages, gender, and religion. Muslims follow specific Islamic guidelines that prohibit them from taking a vaccine with certain ingredients. This study aims at analyzing Facebook and Twitter data to understand the discourse related to halal vaccines using aspect-based sentiment analysis and text emotion analysis. We searched for the term "halal vaccine" and limited the timeline to the period between 1 January 2020, and 30 April 2021, and collected 6037 tweets and 3918 Facebook posts. We performed data preprocessing on tweets and Facebook posts and built the Latent Dirichlet Allocation (LDA) model to identify topics. Calculating the sentiment analysis for each topic was the next step. Finally, this study further investigates emotions in the data using the National Research Council of Canada Emotion Lexicon. Our analysis identified four topics in each of the Twitter dataset and Facebook dataset. Two topics of "COVID-19 vaccine" and "halal vaccine" are shared between the two datasets. The other two topics in tweets are "halal certificate" and "must halal", while "sinovac vaccine" and "ulema council" are two other topics in the Facebook dataset. The sentiment analysis shows that the sentiment toward halal vaccine is mostly neutral in Twitter data, whereas it is positive in Facebook data. The emotion analysis indicates that trust is the most present emotion among the top three emotions in both datasets, followed by anticipation and fear.