Fitbit Analyzer

Analyzer and statistics generator for Fitbit data like sleep, steps count and heart rate. Check out also my related introductory article.

Includes OAuth 2.0 scraper to download personal Fitbit data using the Fitbit API Python Client Implementation.

Usage

Run

 python setup.py install

You will then be able to run the scraper via

fitbit-scraper --id <your_client_id> --secret <your_client_secret> --out dump/destination/folder/path

The client-id and client-secret can be obtained by creating a Fitbit app on the official website. The scraper will than take care to download all data to the given folder, avoiding to re-download data already present (apart for the last present day, which will be re-scraped to avoid missing entries). See fitbit-scraper --help for a detailed usage description.

Data Format

In the util folder you can find some code for loading sleep, heartbeat and steps data as dumped by the scraper script. For the data-dump folder I have one folder per year, and then one sub-folder for each day, in which the different files are generated (e.g. sleep.json, steps.json).

Example:

├───2016
│   ├───2016-01-01
│   │       sleep.json
│   │       steps.json
│   │
│   └───2016-01-02
│           heartbeat.json
│
└───2017
    └───2017-01-11
            calories.json
            distance.json
            elevation.json
            floors.json
            heartbeat.json
            sleep.json
            steps.json

Analyzer and Stats

Compared to basic scraping this phase has more dependencies; see the requirements.txt file and here for a guide on how to manage environments in Conda.

The Sleep Dashboard notebook can provide a quick way to explore your sleep data. See related article and video. I am currently working on notebooks for the other measures, so all feedback is welcome.

Sleep Stats

Sleep values are mapped like this: 0=none (no measure taken), 1=sleeping, 2=restless, 3=awake.

• Basic Stats (sleep values count, sleep efficiency, hours of sleep, total minutes in bed, N max-values for each stat)
• Timing Stats (first minute asleep, to bed time, wake up time, sleep interval min/max length)
• Intervals Stats for each day all the sleep intervals lengths
• Intraday Stats minute to minute report for each day, for the specified sleep values. Total value count, with normalization and centering on specific time.

Heart Rate Stats

• Basic Stats (count, mean, min, max, std) values for the entire dataset, or for aggregated subsets. Common aggregation features are date, year, month, day, day of week and hour
• Daily Stats plotting of average heart-rate by day, month, year, weekday
• Min/Max Values get the N min or max heartbeat values in the dataset

Steps Stats

• Daily Count steps count per day

NEW

Combined Stats

• Correlation correlation coefficient between heterogeneous stats (e.g. steps and sleep)
• Combined TOGGL and heart-rate stats explore relations between Toggl activities and projects and heart-rate

TODO

• Toggl data exploratory notebook
• moving average sleep (days for achieving flatness)
• correlation between constant exercise/running and decreased heart-rate
• formal verification of causality (e.g. good sleep causes to walk, or walk cause good sleep)
• derive sleep-states (REM/NREM) stats (e.g. identify, how long in each) from basic ones, or via correlation with heartbeat (as for now doesn't seem to be feasible actual ECG is still the more proper option)
• max HR and zones (% of MHR)

License

Released under version 2.0 of the Apache License.