Wind estimation from DJI Air 3S hover telemetry: dataset and code

Processed data and analysis code supporting “Horizontal wind estimation from consumer-drone hover telemetry: comparing machine-learning and calibration approaches.” The dataset covers 50 DJI Air 3S hover sessions at two sites in Bremen, Germany, with reference wind measurements from a TriSonica Mini ultrasonic anemometer. The study compares Ridge regression with a tilt-based calibration baseline for estimating horizontal wind from drone telemetry. The archive includes aligned 10 Hz data, derived features, calibration measurements, fitted models, held-out predictions, validation metrics, and processing and plotting scripts. Raw flight and anemometer logs are not included but are available from the corresponding author upon reasonable request.