Closed HagenFritz closed 4 years ago
The issue was that extra days were added to each individual based on the data request from the API. These days were auto-populated with seemingly useless data so I edited the make_dataset.py
script to remove days from the Fitbit daily records when the activities_steps are greater than 0. Even if the participant only wore their fitbit to sleep, their motion should count at least a few steps...
Fixed the issues. Data availability now makes more sense. See the figure here
With the heart rate values processed, we need to remove data from sleep and any other relevant data sources that correspond to poor heart rate readings. This will ensure that we are only getting meaningful values.