Tac-Miner: Visual Tactic Mining for Multiple Table Tennis Matches.

Area of Science:

• Sports Science
• Data Visualization
• Human-Computer Interaction

Background:

• Table tennis tactics, defined by sequences of strokes, are crucial for understanding player strategies and performance.
• Current tactical analysis methods are limited by data volume and dimensionality, hindering comparative and correlational insights.
• Existing research often focuses on single matches, failing to capture dynamic strategy adjustments against different opponents.

Purpose of the Study:

• To develop an effective visual analytics system for analyzing table tennis tactics across multiple matches.
• To address the limitations of traditional statistical charts in handling complex tactical data and enabling comparative analysis.
• To provide a tool for deeper insights into player styles, strengths, and weaknesses by analyzing multi-match data.

Main Methods:

• Developed Tac-Miner, a visual analytics system integrating advanced embedding and dimension reduction algorithms.
• Incorporated an interactive glyph within the system for enhanced data exploration and comparison.
• Utilized multi-match data to enable comprehensive tactical analysis and strategy evaluation.

Main Results:

• Tac-Miner facilitates effective analysis, exploration, and comparison of table tennis tactics from multiple matches.
• The system overcomes the limitations of traditional charts for complex, high-dimensional tactical data.
• User study confirmed the usability of the interactive glyph, and a case study validated the system's utility with expert-approved insights.

Conclusions:

• Tac-Miner offers a novel and effective approach to table tennis tactical analysis, moving beyond single-match limitations.
• The system enhances the ability to uncover player strategies, playing styles, and performance nuances.
• The integration of advanced algorithms and interactive visualization provides valuable tools for coaches and analysts.