Wearable ECG–IMU and respiration-calorimetry data for heat-production prediction in growing pigs

Network-synchronized clocks were used for the ECG, inertial measurement unit (IMU) and calorimetry acquisition computers, allowing records to be aligned by pig identity and timestamp. The planned release contains raw or minimally processed ECG and IMU signals, R-peak and heart-rate-derived data, 5-min calorimetry and multimodal matched records, animal and experimental metadata, the final 30-min analysis panel, variable dictionaries, quality-control and exclusion records, pig-grouped nested-cross-validation predictions, short-term individual-calibration outputs, and source data underlying the main figures and tables.

After quality control, 4,079 matched 30-min windows were retained; the post-adaptation analysis set contained 2,470 windows. The dataset supports research on wearable animal physiology, ECG–IMU sensor fusion, heat-production prediction, subject-grouped machine-learning validation and short-term individual calibration.