Deep Learning–Based Sentiment and Topic Analysis of Turkish Football Fans on X Platform

ABSTRACT This study examines social media fan comments using big data analytics to extract meaningful insights for football clubs. Comments from Turkey's top football clubs—Beşiktaş, Fenerbahçe, Galatasaray, and Trabzonspor—posted on the X platform between May 26, 2024, and November 11, 2024, were analyzed with sentiment analysis and deep learning techniques. The dataset of 20,000 comments was preprocessed and classified based on sentiment using a deep learning model implemented in Python via the X API. Additionally, text mining and topic modeling techniques identified frequently used words and key themes in the comments. The model's performance was evaluated using sensitivity, specificity, accuracy, and F 1 score metrics. Sentiment analysis results demonstrated high performance, with Beşiktaş achieving precision, recall, and F 1 scores of 0.957, 0.941, and 0.949, respectively, while Fenerbahçe scored 0.968, 0.941, and 0.954. Trabzonspor had the highest proportion of negative sentiment at 25.6%, reflecting the most critical fanbase. These findings highlight the effectiveness of deep learning and sentiment analysis in assessing fan engagement on social media, offering valuable insights for football clubs to better understand supporter sentiment and refine their strategies.