Description
Numerous challenges exist when trying to detect changes in a person’s health status using data from wearable sensors. A core issue is that wearable sensor data varies tremendously between individuals and over time due to a person’s daily activities. “Reasonably comparable” time periods must be identified based on activity levels and other variables, which can then be analyzed to determine if a change in health status has occurred. SciPy, NumPy, Pandas, and Matplotlib were used to process the massive number of variable combinations in our data and to output visualizations that highlight useful patterns in the data.