A network view of social media platform history: Social structure, dynamics and content on YouTube

Abstract

Social media sites are prone to change from many internal and external causes, yet it is difficult to directly explore their histories in terms of the content itself. Search and browsing features are biased toward new and paid content, archives are difficult to navigate systematically, and their scale makes any observations challenging to contextualize. Here, we present results of an ongoing study of YouTube's history (currently with more than 76 million videos) using a combination of iterative browsing, network crawling and clustering within and across time periods. Through this method, we are able to identify historical patterns in YouTube's content related to internal and external events. Our approach thus illustrates an adaptation of network analysis for understanding the content histories of social media platforms.