Scikit Digital Health (SKDH) is a Python package with methods for ingesting and analyzing wearable inertial sensor data.

• Documentation: https://scikit-digital-health.readthedocs.io/en/latest/
• Bug reports: https://github.com/PfizerRD/scikit-digital-health/issues
• Contributing: https://scikit-digital-health.readthedocs.io/en/latest/src/dev/contributing.html

SKDH provides the following:

• Methods for ingesting data from binary file formats (ie Axivity, GeneActiv)
• Preprocessing of accelerometer data
• Common time-series signal features
• Common time-series/inertial data analysis functions
• Inertial data analysis algorithms (ie gait, sit-to-stand, sleep, activity)

Availability

SKDH is available on both conda-forge and PyPI.

conda install scikit-digital-health -c conda-forge

or

pip install scikit-digital-health

Warning

Windows pre-built wheels are provided as-is, with limited/no testing on changes made to compile extensions for Windows.

Note

Windows users may need to install an additional requirement: Microsoft Visual C++ redistributable >14.0. The 2015 version can be found here: https://www.microsoft.com/en-us/download/details.aspx?id=53587

Build Requirements

As of 0.9.15, Scikit Digital Health is built using Meson.

Citation

If you use SKDH in your research, please include the following citation:

[1] L. Adamowicz, Y. Christakis, M. D. Czech, and T. Adamusiak, “SciKit Digital Health: Python Package for Streamlined Wearable Inertial Sensor Data Processing,” JMIR mHealth and uHealth, vol. 10, no. 4, p. e36762, Apr. 2022, doi: 10.2196/36762.