Single-camera markerless gait analysis in Parkinson's disease — data and analysis code

Data and analysis code accompanying the article Single-camera markerless gait analysis in Parkinson's disease: validation and levodopa response. Running the two scripts in code/ reproduces every number reported in Tables 1 and 2, in the Abstract and Results, and every value shown in Figures 2 and 3 and in Supplementary Figure S1. Contents: data/ — 1,163 strides from 10 healthy adults (per stride, per recording source: front video, side video, marker-based motion capture); measured and reconstructed standing height; per-participant summary stride lengths for the scale analysis; and gait parameters of 73 patients with Parkinson's disease in both medication states under each trajectory source, with the MDS-UPDRS III gait item. code/ — study1_validation.py (agreement with motion capture, equivalence testing, measurement precision, scale analysis) and study2_levodopa.py (levodopa response, robustness across trajectory sources, comparison with the clinical gait rating). RESULTS.txt — the output of both scripts. Requires Python 3 with numpy, pandas and scipy. Participants are pseudonymous (C01–C10, P01–P73); no key linking labels to individuals is distributed. See README.md inside the archive for full column-level documentation.