Data gap handling is how Unfair treats missing days in biomarker, dose, and subjective-log streams. Most users miss some data — a wearable not charged, a travel day without logging, a forgotten morning check-in — and how those gaps are handled quietly shapes every chart and every ranked recommendation.
Why it matters
Two wrong approaches are common in other tools:
- Silent interpolation. Filling a missing night with a copy of the previous night makes charts look clean and hides the fact that the user has no real data. This pollutes downstream analysis.
- Silent exclusion. Dropping missing days from a moving average window without noting it can make a 7-day average computed on 3 actual days look misleadingly stable.
Unfair uses explicit, visible gap handling instead — charts mark gaps, averages note their actual denominator, and confidence weights drop when gaps are large enough to matter.
How Unfair treats gaps
The basic rules:
- Missed biomarker day — charted as a gap with a visible break in the line; excluded from the rolling average; the "days used in average" label updates.
- Missed dose log — logged as a skipped dose if reminders were sent; adherence and consistency score reflect it.
- Missed subjective check-in — logged as missing; if gaps exceed 3 in a 7-day window, the confidence on all subjective charts drops visibly and the feedback loop holds rank changes.
- Multiple consecutive gap days (> 3) — Unfair prompts the user to confirm whether the gap was a sync issue, a deliberate pause, or a logging lapse. This disambiguates the three into different states that affect recommendations differently.
Why gap marking matters for trust
A chart that shows a clean line can convince a user of a trend that does not exist in the data. A chart that shows real gaps, even if it is uglier, produces better decisions. The ranked output inherits this discipline — recommendations are held or downgraded in confidence when the source data is sparse, rather than confidently drifting based on half-complete logs.
Gap patterns worth investigating
Some gap patterns are themselves diagnostic:
- Morning logging gaps only — usually an onboarding or habit issue.
- Weekend-only gaps — lifestyle drift on non-work days; worth distinguishing from weekday data.
- Recurring multi-day gaps every 3–4 weeks — often travel; worth tagging explicitly.
- Gaps that coincide with stack changes — a quiet sign that the user is avoiding logging when they expect a bad result.
Unfair does not moralize about gaps; it marks them and adjusts confidence.
How this appears in Unfair
Gap markers are visible on every trend chart as broken lines or greyed segments, and averages display the actual count of days included. When gaps accumulate past the confidence threshold, the review screen surfaces a "data coverage is thin" note rather than silently producing a result the user will over-trust.
Clinical safety note
A long gap between a new compound start and the next log is itself worth noting — particularly if a safety signal would have surfaced in the logs. When gaps coincide with new compounds, catch up on logging before drawing conclusions about tolerance or side effects.