New acceleration table in DB

[watsonix]

This will be a large amount of data (to get activity from acceleration we need many accelerometer readings). Point is eventually to exclude intense activity from HRV profiles. (body activity increases heart rate and distorts HRV readings)

• Figure out how to store efficiently for later retrieval?
• Should have user, mobile_time and acceleration (potentially multiple values) fields

[watsonix]

One simple way to get "activity" from acceleration readings is to just take the standard deviation of many acceleration values. Acceleration values should be sampled at the rate of 10s of milliseconds

[ayumi]

For reference: https://cordova.apache.org/docs/en/latest/reference/cordova-plugin-device-motion/

Properties
x: Amount of acceleration on the x-axis. (in m/s^2) (Number)
y: Amount of acceleration on the y-axis. (in m/s^2) (Number)
z: Amount of acceleration on the z-axis. (in m/s^2) (Number)
timestamp: Creation timestamp in milliseconds. (DOMTimeStamp)

[ayumi]

• recording 10 measurements /sec for 16 hrs/day == 576,000 row/day /user (which is a lot)

• suggestion: app samples 10 accelerometer readings /sec, then it calculates the average and uploads records at a slower interval – maybe an interval similar to the HRV uploads.

• user_id

• mobileTime (time of upload / end of interval)

• x_average

• y_average

• z_average