Public opinion and emotional discourse: a study of YouTube comments on Mexican news in 2024 Academic Article in Scopus uri icon

abstract

  • In recent years, integrating social media data and natural language processing techniques to handle large volumes of data has become more popular, particularly for gaining insights into user behavior in digital environments. Unlike traditional methods such as surveys, using data from social media has allowed for capturing broader and more spontaneous dynamics. In this context, YouTube stands out as a particularly relevant platform in Mexico, the fifth country in the world in terms of users and where this platform is the second main source for news consumption. This study introduces a dataset comprising 1,048,725 comments extracted from 14,161 news videos published in 2024 by eight national channels on YouTube. We aim to examine the public discourse among the Mexican audience based on the extracted data. We applied sentiment and emotion analysis techniques, as well as hate speech and aggressiveness detection, and utilized topic modeling with BERTopic. Likewise, we consider theme polarization by evaluating the distribution of categories in sentiment, hate, and aggression. The results reveal that the prevailing content is about politics, and nearly 45% of comments are posted on the same day a news video is released, indicating that most activity occurs on the day of publication. Moreover, topics related to judicial reform, corruption, violence, and criticism of the government are the most frequent, and they are dominated by negative reactions. Although topic polarization is low, there is a clear tendency toward negative stances. By mapping the emotional tone of digital discourse in Mexico, this study identifies the primary areas of citizen concern and sheds light on the prevailing public sentiment regarding current affairs. © The Author(s), under exclusive licence to Springer Nature Switzerland AG 2025.

publication date

  • June 1, 2026