reframing the same body
summary
an experiment in how easily data can be misconstrued: one year of wearable data, framed two opposite ways.
the experiment
how easily can the same data be made to say opposite things? this was a visualization exercise rather than a deep analytics project: I took a year of biometric data from my ultrahuman ring (recovery score, sleep score, step count, jan–nov 2025), left every number alone, and changed only the design choices around them: ordering, scale, color and the treatment of missing data.
experiment
nine months of my step counts. scroll to apply four design choices, one at a time. the numbers never change, only the way they're shown.
- 01missing monthslabeled → dropped
- 02ordercalendar → ascending
- 03colorwarning → reward
- 04emphasisthe low → the peak
- reads asdeclineprogress
- values changed0
monthly step totals, jan–nov 2025, from my ultrahuman ring (july and august have no data). same numbers as the tableau dashboards below.
frame a: decline
shown chronologically, recovery peaks early and drops sharply through the summer. missing data is labeled, breaking the line and drawing attention to absence. steps trend down and end on a labeled low. the year reads as interrupted and flat.
frame b: progress
the exact same values, reordered into an upward progression: recovery climbs in green, steps rise to a large labeled peak, sleep bars build from short to tall, and gaps are handled quietly. the same body now reads as disciplined and improving.
why it matters
nothing about the data changed, only the visual logic. wearables are marketed as self-improvement tools, but they rely on visualization to tell that story. data doesn't speak for itself; it's always arranged, filtered and framed.