Records
358,497
daily observations
01 / Analytical platform
Movement,
made legible.
A measured view of 358,497 fitness records across six months, turning daily signals into patterns you can inspect.
The collection
Apr-Sep
2023
Users1,959Activities07StorageParquet
At a glance
The first read
Users
1,959
distinct profiles
Average steps
6,348
per record
Average calories
754
burned per record
What is inside
A fixed cohort, made inspectable.
The dataset covers 182 days, 7 activity types, and a consistent set of movement and recovery signals. Every view is derived from the same exported source.
01
Movement
Steps, calories, and activity mix over time.
02
Recovery
Sleep and heart-rate signals placed beside activity.
03
Models
Classifier accuracy: 89.3%.
04
Quality
Exported checks keep the analytical surface honest.
Explore patterns
Compare activity, health, time, and user behavior.
Inspect quality
See completeness, ranges, outliers, and findings.
ƒ(x)Test a profile
Run an instant client-side baseline prediction.