Understanding digital ethnography: Socio-computational analysis of trending youtube videos

YouTube, the online video sharing website, was launched in February 2005to help people host numerous amateur and professional videos. Since then, YouTube hasrapidly grown to be a cultural phenomenon for its massive user base. Still there is a lackof systematic research focusing on the video-based social networking platformcompared to other social media platforms. In this research, we study top 200 YouTubevideos trending daily for a 40-day period separately in the United States of America(USA) and the Great Britain region (GB), resulting in close to a 8000-video dataset foreach region. The dataset includes title, category, URL, video ID, comments, views, likes,and dislikes for the videos. The dataset is further enhanced by extracting associationswith other social media platforms of the YouTube channels that are hosting thesetrending videos. Some of the findings from analyzing this dataset include, (1) the morecomments a video has, the more neutral the sentiments are, and videos that have fewercomments have more polarized sentiments; (2) similar behavior is observed when thenumber of views is correlated with sentiments of comments; (3) videos that wereassociated with 10 online platforms (including a wide variety of social media platforms)received the maximum number of views-any more or less did not help in increasing theviews. This study sheds light on the digital ethnographic behaviors in terms of videobased content generation, sharing, and consumption, and further allows us to gleansimilarities and differences between USA and GB regions.