Topic modelling of racist and xenophobic YouTube comments. Analyzing hate speech against migrants and refugees spread through YouTube in Spanish
Juan Pablo Latorre, Javier J. Amores
Abstract
Racist and xenophobic hate speech transmitted online is of increasing concern, as social media provide the most suitable channel for its rapid, uncontrolled and massive spread. However, although this type of hate has been studied on diverse social platforms, the academy has not yet focused the attention on the speech generated on YouTube. This work aims to fill this gap in the literature, analyzing the racist and xenophobic hate speech that is spread through YouTube comments in Spanish. Specifically, a manual classification has been carried out to detect hate messages, as well as latent sentiments in comments. On the other hand, a topic modelling was carried out exclusively on hate comments. The results show the characteristics of the racist/xenophobic discourse that spreads through YouTube in Spanish, which, in general terms, represents migration as a burden and a threat, and calls for a government that carries out tougher and more restrictive immigration measures.