Sentiment analysis of the 2024 Paris Olympics using RoBERTa language model

Purpose

This study investigates public views of the Paris 2024 Olympics by analyzing tweets with the aim of elucidating public perceptions and identifying the factors that influence engagement, sentiment, and discourse related to this mega event.
Design/methodology/approach

Two advanced models, ChatGPT and Roberta, were used to compare sentiment analysis on the dataset, which included 30,300 valid tweets. This research analyzed by RoBERTa the trends and fluctuations of the sentiment in engagement, regarding cultural, social, and geopolitical aspects. Sentiment trends were analyzed by country, sport, gender, user engagement, etc.
Findings

91.1% of the tweets were either neutral or positive, showing a generally favorable public opinion. Certain events, such as the basketball semifinals and the javelin finals, received high engagement, while the opening ceremony garnered mixed reactions, underscoring the importance of cultural and social sensitivities in event design. Country-specific analysis showed positive feedback from the USA and India, national pride and athletic success, and negative sentiment from Russia, possibly because of the country’s exclusion from this Olympics. The top three sports with positive reactions were basketball, athletics, and swimming. On the other hand, boxing was criticized for fairness debates. A gender-based sentiment analysis showed positive sentiment toward female athletes in 57% of the tweets. Negative tweets, although lesser in number, received more likes and retweets.
Originality/value

This paper investigates sentiment fluctuations during the Paris 2024 Olympics using advanced language models, exploring the convergence of sports, culture, and public perception to provide actionable insights.