search

beeminder / BeeSwift

Official Beeminder for iOS app
Other
30 stars 6 forks source link

Inaccurate Mindful Minutes #452

Closed sachiesaaan closed 7 months ago

sachiesaaan commented 7 months ago

Replicata

  1. Open Health app integration
  2. Add Mindful Minutes as source
  3. Add Data to Apple Health
  4. Check Data on Beeminder

Expectata

data as itself

2min →　2min

Resultata

every data is counted as +1

ex) 2min →3min, 10min →11min

ex2) If I added 2min and 1min in a day, it is counted as 2+1 min + 1+1min = 5min

Nota Nebulosa

IMG_4273 IMG_4274

Verbata: no need, to add, keywords, when, first creating, a gissue,

theospears commented 7 months ago

Thanks for the thorough bug report!

The current beeminder algorithm for this metric is: +1 minute for every wall clock minute contained within a meditation session So e.g. a session from 6:00:30 - 6:01:15 will count as 2 min. Note this is slightly different to the reported behavior - e.g. a session which is 6:00:05 - 6:00:55 will be reported correctly.

Based on some synthetic test data, the Apple Health algorithm appears to be Calculate the total duration of all meditation sessions, without double-counting overlap, and round to the nearest minute, with 30s rounding up So e.g. a pair of sessions 6:00:03 - 6:00:23 (20s) and 6:00:01 - 6:00:21 (20s) will total 22 seconds, and round to 0 minutes.

For Beeminder we could either switch to exactly matching the apple algorithm, or leave out the rounding so we can report non-integer numbers of minutes. Curious if people have thoughts on which is better.

sachiesaaan commented 6 months ago

Thank you for fixing this! Personally, I do sessions for some minutes and never care about seconds so rounding to the nearest minute is fine..