@Technonium@WHOOP If it runs off recovery, a sick week is going to be seven grim planets in a row. A month of these side by side would make a weirdly honest chart.
The Doctor Visit Pack tool reads your Apple Health export in the browser. The file is never uploaded, and the page runs no analytics.
You get a 2-page PDF to bring to your appointment.
https://t.co/9eoyBwuGLI
Apple Health stores HRV as SDNN, from short spot readings. WHOOP's API returns hrv_rmssd_milli: RMSSD, averaged across sleep. Same label, different statistic, different window. If you hand both to an assistant, tell it that, or trend each device only against its own baseline.
A sleep export has many rows per night, not one. Each row is a stage segment with a start, an end and a state.
For total sleep, sum the asleep segments for that night. The guide lists the stages.
https://t.co/56dZaCfSvW
@alexmichael The bit I'd check before opting in: questions get de-identified before they reach OpenEvidence, and it deletes the question and answer data within 24 hours.
@sanyayyyy High and low only mean something against your own baseline on one device. Apple Watch stores SDNN from short spot readings while WHOOP and Oura use overnight RMSSD, so the same night can look very different depending on what's on your wrist.
@_StreetWise21 Same legs, eight and a half minutes apart over 42.2k. Garmin starts from its VO2 max estimate, Strava from the runs you've actually logged. A hard half in the next few weeks would settle which one is closer.
@luca_iaconelli@WHOOP Five tennis days a week will include some late matches, and a hard evening session tends to pull that night's HRV down. So some low-recovery mornings are the match showing up, not bad sleep.
@SanderBelaen The sensors overlap, the HRV doesn't quite: Apple Health stores SDNN from a handful of short readings while Whoop scores overnight RMSSD, so a Recovery 91 built on the Watch is a different recipe.
@vishalskotak If Haiku wrote the practice chats from clean data, real exports will throw overlapping sleep at it: the Watch and a second sleep app both logging the same night, which a 0.8B model will happily add up.
@the5krunner Did the awful-night signal hold up for people with less than 90 days of history, or does a new watch spend its first three months guessing?
@_weswen Try asking it about a night you flew across time zones. Sleep that starts on one date and ends on another is where "last night" gets slippery, and the answer depends on which timezone n8n buckets by.
@GodsBoy7777 The harder case is a day that exists but is only half there: the Watch died at 3pm, so steps come back as a real-looking 4,000 instead of date_unavailable.
Two devices recording the same thing get two rows in the Doctor Visit Pack, never an average.
Steps are the exception: the highest single-device total each day, because iPhone and Apple Watch count the same steps.
https://t.co/pfDc3Ph3Ht
Observatory 1.2.0 has a Home Assistant pack: three blueprints and a plain-cards Body dashboard.
Wind-down sleep length, a recovery-aware morning, a low-data warning. The guide compares routes.
https://t.co/FX3FqfFwpt
Switching to Android? Your Apple Health history stays on the iPhone. iCloud syncs it between Apple devices only.
Make a CSV or JSON copy while the iPhone is still set up, and keep it somewhere you control.
https://t.co/ok47szstfC
@Greg_TheBuilder If the step total looks suspiciously high, it's usually the iPhone and the Watch both logging the same walk in export.xml. The Health app only counts one of them, a raw sum counts both.
@kottley@Musecases@Muse Apple Health only stores HRV as SDNN from short Watch readings, and Garmin's HRV Status runs on overnight RMSSD, so that tile is a different metric wearing Garmin's name.