Portal de Periódicos da CAPES

Application of social media data mining in biomechanical and tactical analysis of tennis tournament players

2024; Volume: 21; Issue: 3 Linguagem: Inglês

10.62617/mcb425

1556-5300

Hongmin Yu, Xing Wei,

E-commerce and Technology Innovations

The rise of social media has provided a rich source of real-time data for analyzing player performance and tactics in professional sports, particularly tennis. This study harnesses social media data mining techniques to analyze tennis-related discussions on Twitter, focusing on identifying biomechanical patterns and tactical strategies during major tournaments. We propose a hybrid model combining Bidirectional Encoder Representations from Transformers (BERT) for generating contextual embeddings and Bidirectional Long Short-Term Memory (Bi-LSTM) for analyzing the sequential nature of tweets. The data collection spans tweets discussing key tournaments, including the Australian Open, French Open, Wimbledon, and US Open. It focuses on specific player movements such as footwork, speed, endurance, and tactical decisions like serve placement, net play, and shot selection. Our methodology includes preprocessing the data, tokenizing the text, and applying sentiment analysis to capture public perception of player performance. The model achieves an accuracy of 88.5% and an F1-score of 87.95%, outperforming comparative models such as BERT with CNN and GloVe with LSTM. The analysis highlights key player-specific tactics, including Rafael Nadal’s baseline dominance and Novak Djokovic’s defensive play, as well as tournament-specific strategies, such as serve-and-volley at Wimbledon and baseline control at the French Open. Furthermore, sentiment analysis reveals positive public perception toward player performance, with key emotions such as excitement and admiration frequently expressed during intense match moments. This study demonstrates the effectiveness of applying advanced NLP techniques to social media data for sports analytics. The insights generated can inform players, coaches, and analysts in enhancing performance strategies and understanding public reactions. Using social media data, our approach provides a scalable framework for analyzing tactical shifts and player performance in other sports contexts.