Anti-Stress-Tracker

Project for GCT501 course at KAIST. It aims to provide user-friendly stress estimations to help people manage their stress. It works with Fitbit Web API and need heart rate data to operate. For testing purposes authours used Fitbit Charge 2 wristband.

Running the app

Dependencies:

Flask (http://flask.pocoo.org/docs/1.0/quickstart/), template renderer: (http://jinja.pocoo.org/docs/2.10/)
python-fitbit (https://github.com/orcasgit/python-fitbit).
config.ini file should be prepared and put into project directory (see next session).

Starting the server

export FLASK_ENV=development
export FLASK_APP=init.py
flask run

These commands should be run from the root of the project directory. Server running in development mode automatically restarts when code changes.

Accessing your Fitbit data

To make the application work with you personal records from your Fitbit, you need

Have a Fitbit wristband, that can measure heart rate and Fitbit acccount.
Register an new app on Fitbit web-site
- Login into your account on fitbit.com
- Go to https://dev.fitbit.com/apps/new
- Register a new app of PERSONAL type, other parameters are not significant
- You will receive Client ID and Client Secret of the app
Receive ACCESS_TOKEN, REFRESH_TOKEN
- From homepage of you app on dev.fitbit.com (with Client ID info) go to OAuth 2.0 tutorial page (link at the end of the page)
- Choose Authorization Code Flow
- You should allow at least heart rate, sleep and workout scopes
- Perform steps 1, 1A and 2 to get your access token and refresh token
At the root of the project directory create config.ini text file of the following format and fill the blanks with the data you got. Expires at fill with any int number, e.g. 3600.

config.ini format

[Login Parameters]
c_id = 
c_secret = 
u_id = 
access_token = 
refresh_token = 
expires_at =

You are all set, now you can run the server and see you stress level at http://127.0.0.1/5000

Stress estimation

Estimation of the stress level is based on the Baevsky Stress index formula, that estimates stress level using heart rate variability. For reference see http://www.cardiometry.net/issues/no10-may-2017/heart-rate-variability-analysis. Originally from R.M. Bayevcky, O.I. Kirillov, S.Z. Kletskin. Mathematical analysis of Cardiac Rhythm under stress. In Hauka, 1984. [In Russian]

Formula

sress_score = AMo / (2 * VR * Mo), 
where
R-R interval - time delay between consecutive heart beats (assumed to be inverse value of the bpm)
Mo - mode - most frequent R-R interval value
AMo - mode amplitude - % of the intervals corresponding to Mode
VR - variational range - difference  between min and max R-R intervals

maria-korosteleva/Anti-Stress-Tracker