SleepPy: A python package for sleep analysis from accelerometer data

Measures of sleep quality and quantity can provide valuable insights into the health and well-being of an individual. Traditionally, sleep assessments are performed in the clinic/hospital setting using polysomnography tests. Recent advances in wearable sensor technology have enabled objective assessment of sleep at home. Actigraphy has been widely used for this purpose and several algorithms have been published in the literature over the years. However, implementation of these algorithms is not widely available, which creates a barrier for wider adoption of wearable devices in clinical research.

SleepPy is an open source python package incorporating several published algorithms in a modular framework and providing a suite of measures for the assessment of sleep quantity and quality. The package can process multi-day streams of raw accelerometer data (X, Y & Z) from wrist-worn wearable devices to produce sleep reports and visualizations for each recording day (24-hour period). The reports are formatted to facilitate statistical analysis of sleep measures. Visualization acts as a quick debugging tool, provides insights into sleep patterns of individual subjects and can be used for presentation of data to diverse audiences.