The MultiSenseBadminton dataset is a comprehensive, multi-sensor dataset for evaluating badminton performance, featuring 7,763 badminton swing data from 25 players. It includes sensor data on eye tracking, body tracking, muscle signals, and foot pressure, along with video recordings, detailed annotations on stroke type, skill level, sound, ball landing, and hitting location, as well as survey and interview data. The dataset was developed based on insights from interviews with badminton coaches to ensure usability and comprehensiveness for training and research. It covers various skill levels, including beginners, intermediates, and experts, and provides resources for understanding biomechanics across skill levels. The dataset includes five types of sensor data streams, along with expert interviews, surveys, and annotated data. It also includes video recordings from different viewpoints and detailed annotations for stroke quality evaluation. The dataset was validated using a proof-of-concept machine learning model, demonstrating its applicability in advanced badminton training and research. The dataset is publicly available on figshare and includes physiological data such as EMG and gaze, behavioral data such as foot pressure and joint movement, as well as video data for sensor visualization and annotation data. The dataset is organized in a hierarchical structure, with each subject having sensor-data HDF5 files labeled with the date of data collection and the participant ID. The HDF5 files can be accessed using HDF5 viewer software, and the Python code for reading these files is available on the project's GitHub repository. The dataset includes detailed annotations for stroke types, skill levels, landing positions, hitting locations, and hitting sounds, providing a comprehensive representation of badminton strokes and related characteristics. The dataset is designed to support training programs, performance analysis techniques, and coaching strategies in badminton. The dataset was collected using a shuttlecock launcher and three cameras to capture front, side, and full views of the participants, ensuring consistent data collection. The data collection process involved calibration, instructional videos, and practice strokes, with real-time monitoring and annotation of sensor data. The dataset includes data summary files and survey data, along with sensor data, providing a detailed understanding of participants' characteristics and study findings. The dataset is structured with a tree-like organization, with the top-level folder being an archive folder containing four types of files, a data-summary file, an interview file, an annotation data file, and a survey-data file, along with subject folders. The dataset is publicly available and includes data on participants' physical attributes, training experience, and subjective experiences with wearing sensors. The dataset is designed to support the development of training systems and research in badminton, providing a comprehensive and detailed representation of badminton strokes and related characteristics.The MultiSenseBadminton dataset is a comprehensive, multi-sensor dataset for evaluating badminton performance, featuring 7,763 badminton swing data from 25 players. It includes sensor data on eye tracking, body tracking, muscle signals, and foot pressure, along with video recordings, detailed annotations on stroke type, skill level, sound, ball landing, and hitting location, as well as survey and interview data. The dataset was developed based on insights from interviews with badminton coaches to ensure usability and comprehensiveness for training and research. It covers various skill levels, including beginners, intermediates, and experts, and provides resources for understanding biomechanics across skill levels. The dataset includes five types of sensor data streams, along with expert interviews, surveys, and annotated data. It also includes video recordings from different viewpoints and detailed annotations for stroke quality evaluation. The dataset was validated using a proof-of-concept machine learning model, demonstrating its applicability in advanced badminton training and research. The dataset is publicly available on figshare and includes physiological data such as EMG and gaze, behavioral data such as foot pressure and joint movement, as well as video data for sensor visualization and annotation data. The dataset is organized in a hierarchical structure, with each subject having sensor-data HDF5 files labeled with the date of data collection and the participant ID. The HDF5 files can be accessed using HDF5 viewer software, and the Python code for reading these files is available on the project's GitHub repository. The dataset includes detailed annotations for stroke types, skill levels, landing positions, hitting locations, and hitting sounds, providing a comprehensive representation of badminton strokes and related characteristics. The dataset is designed to support training programs, performance analysis techniques, and coaching strategies in badminton. The dataset was collected using a shuttlecock launcher and three cameras to capture front, side, and full views of the participants, ensuring consistent data collection. The data collection process involved calibration, instructional videos, and practice strokes, with real-time monitoring and annotation of sensor data. The dataset includes data summary files and survey data, along with sensor data, providing a detailed understanding of participants' characteristics and study findings. The dataset is structured with a tree-like organization, with the top-level folder being an archive folder containing four types of files, a data-summary file, an interview file, an annotation data file, and a survey-data file, along with subject folders. The dataset is publicly available and includes data on participants' physical attributes, training experience, and subjective experiences with wearing sensors. The dataset is designed to support the development of training systems and research in badminton, providing a comprehensive and detailed representation of badminton strokes and related characteristics.