The International Affiliation Network of YouTube Trends

Abstract

Online video, a ubiquitous, visual, and highly share- able medium, is well-suited to crossing geographic, cultural, and linguistic barriers. Trending videos in particular, by virtue of reaching a large number of viewers in a short span of time, are powerful as both influencers and indicators of international communication flows. In this work, we study a large set of videos trend- ing across 57 nations, collected from YouTube over a 7-month period. We consider the set as a network of content flowing between nations, then develop conditional co-affiliation, a nation-nation co-affiliation index that enables a meaningful interpretation of network path length and the application of betweenness centrality. We observe a highly-interlinked network with remarkably similar co-affiliation levels between very different nations. However, Arabic-speaking nations appear more isolated, with the U.A.E. emerging as a key bridge. By analyzing video trend lifespans, we show that nations having many globally-popular video trends are reliably not the nation where those trends are strongest: we see no evidence to support the widely discussed idea of cultural exporter or trendsetter nations. We model correlations between co-affiliation and a selection of contextual factors. We note a surprisingly complex interaction between migration and shared video trends. Consistent with existing work on video popularity, we find that long trending times within one nation do not necessarily translate to reaching a wide global audience. This work expands on previous studies of the geographic popularity of videos by incorporating trending data and extend- ing our analysis from video-nation affiliations to nation- nation co-affiliations. Characterizing these relationships is key to understanding the international cultural impact and potential of online video.