This consulting project was part of my graduate consulting curriculum, where students provide statistical consulting for the statistics department.
The client (Elsa Bates) was a BMED M.S. student working on their thesis project involving using wearable posture-monitoring sensors (IMUs) to quantify spinal angles, and compare these measurements to the gold standard motion capture system sensors. Elsa needed help determining whether the IMU data is “close enough” to the motion capture data, and how to best summarise and present these findings across all 30 participants in her results section.
In summary, my group did the following:
Data Management (combine files, align times, calculate curve characteristics)
Compare curves wholistically (visually, correlation coefficients, dynamic time warping)
Compare characteristics of curves using linear mixed models (e.g., proportion of time spent in harmful posture, range of motion, etc.)