Public Opinion Toward Viral YouTube Content: Machine Learning–Based Sentiment Analysis of Comments on the MrBeast Channel

Abstract

Social media, particularly YouTube, has become a primary space for the public to express opinions and respond to various forms of digital content. User comments on viral videos can represent audience perceptions and sentiments; Therefore, analyzing comment data is essential to more systematically understand trends in public opinion. This study aims to examine public opinion toward viral content on the MrBeast YouTube channel through a machine learning–based sentiment analysis approach. The data were obtained from a MrBeast YouTube comments dataset available on Kaggle. The dataset was processed through text preprocessing stages, including data cleaning, text normalization, and the removal of irrelevant words. The text was then transformed into numerical features using Term Frequency–Inverse Document Frequency (TF-IDF), before sentiment classification was performed using machine learning algorithms to categorize comments into three classes: positive, neutral, and negative. The results show that the sentiment distribution consists of 63.09% positive sentiment, 35.50% neutral sentiment, and 1.41% negative sentiment. These findings indicate that the majority of audience comments on MrBeast's content tend to be positive, reflecting strong public acceptance of viral content on the channel. This study is expected to contribute to the development of data-driven public opinion analysis on social media and to serve as a reference for future research in sentiment analysis, digital communication, and viral content studies.