Unsupervised Sentiment Analysis of the “Dune: Part Two” Trailer Comments

This research paper explores applying unsupervised sentiment analysis techniques to analyze comments on the official trailer of 'Dune: Part Two‘. The goal is to uncover the underlying sentiments of viewers towards the upcoming movie, without relying on pre-labeled data. Various text preprocessing methods, such as removing accented characters, expanding contractions, and lemmatization, are examined for their impact on sentiment analysis outcomes. Additionally, the study investigates how integrating domain-specific knowledge, like key Dune terminology, influences sentiment analysis results. The Valence Aware Dictionary and sEntiment Reasoner (VADER) tool is utilized for sentiment analysis, offering a lexicon and rule-based approach tailored for social media text. The findings contribute insights into audience sentiment towards the movie, informing marketing and audience engagement strategies. The research highlights the significance of utilizing domain expertise along with text preprocessing for precise sentiment analysis in specific settings like fan communities.