This repository holds the data and the Python data-analysis utilities for a collaborative research effort conducted in the Physical Medicine and Rehabilitation group, contributors include: Dr. Amir Ahmadian, Dr. William Lian, Dr. Jaclyn Nguyen, Dr. Shannon B. Juengst, Dr. Ugo Bitussi, James Kennington, Dr. Fatma Gul, and Dr. Kathleen R Bell. Publication currently under review.
The data were originally contained in an excel spreadsheet in a pivoted format, where each row
represented a day and the columns represented successive intraday measurements of pain scores or
opioid consumption. This raw format is still available in data/MPOW-Data-20190206.xlsx.
The data were extracted from the spreadsheet and normalized into relational form, in which
each row represents a single measurement (either daily or intraday depending on the quantity
being measured). The normalization and extraction utilities are available in mpow/load_data.py,
and the normalized data have been saved to data/data.h5 in HDF5 format.
Though the regression helper functions are in mpow/regression.py, the bulk of the analysis
is in the Jupyter notebook analysis.ipynb. This notebook can be accessed by
running jupyter notebook in a properly-setup python console, or by clicking the binder link
here.