Classifying YouTube channels: Algorithmic approach to channel mapping

YouTube is one of the most influential platforms for news and political commentary; however, its vast scale makes it challenging to study the full range of content and communities that it hosts. Most existing research focuses only on the most popular channels, overlooking the large number of smaller creators who may reach sizable and politically engaged audiences. In this study, we introduce a new approach to identifying and classifying YouTube channels based on patterns in who subscribes to them. Using this method, we mapped the landscape of English-language political channels on YouTube during 2020. Our findings indicate that prior studies have significantly underestimated the prevalence of right-leaning and conspiracy-themed content, which is more prevalent among smaller channels. We also show that large media outlets tend to dominate left-leaning content, while right-leaning content is primarily produced by independent creators. This work demonstrates the value of scalable, data-driven methods for understanding online media ecosystems and offers a more complete picture of the political content available to YouTube users.