Your latest readings against your own usual ranges, and the one thing most worth a look.
Week in review
A day in the life
Typical = median times from the data in the window.
Plan = yours to set; edits save on this device.
Sleep architecture
Trends
Bold = 7-day average; dots = days actually logged. Ticks under
the axis mark logged days. 4+ measures split into one panel each.
Against your baseline
Each selected measure vs your own usual range (shaded =
middle 50% of your history). Green points sit in range; red points are unusual
for you — not necessarily abnormal.
By day of week
Period comparison
Screening test — within-metric permutation of group labels; autocorrelation may inflate significance.
What lines up with what
Habit on one day vs measure N days later, on actively-logged
days only. Orange = rise together, blue = opposite; grey values are consistent with
chance after multiple-comparison control. Click a cell to inspect.
1 day(s)
What precedes your good vs bad days
Pick an outcome; days split into your best and worst third.
For each habit, its level the day before — best-day vs worst-day — sorted by the gap.
Descriptive contrast of your own days, not a causal estimate.
One pair
If this pair was opened because it looked strong above,
read its p as descriptive, not confirmatory.
1 day(s)
Correlation at every delay — click a bar to set it.
Association stability over time
60-day window stepped weekly; hollow points have fewer than 20 paired days. Use this to check whether the association above is stable, drifting, or an artifact of one period.
Before & after
Average course around flagged days; dashed = overall average.
Test a hunch
What predicts good days
Joint description of rank-transformed variables — not a causal model or prediction.
Logging coverage
Days with an actual entry, per measure — imputed zeros don't count.
Goals
Targets set with your provider, scored against your own data over time.
Green = met, against your personal range.
Advisor
Lifestyle associations from the penalized dose-response model —
observational, not causal. Badges: Consistent = three estimators agree
on direction (triangulated association); Assoc. = associational only.
The Confidence column flags thin data, intervals spanning zero, or prior-dominated estimates.