Heart rate
Demand without cause
An elevated value cannot identify whether the context was useful skating, pain-limited walking, heat, waiting, or transportation stress.
WEARABLE HEALTH AND MOBILITY INTELLIGENCE
Context Is the Missing Sensor
Wearables capture signals. The harder problem is preserving what the person was doing, how the body was moving, how transportation affected it, and what the signal can actually support.
This individual case study connects longitudinal wearables, routes, targeted episodic sensors, movement cohorts, transportation context, documentary history, and human review without treating them as interchangeable evidence.
The Context Problem
A Signal Without Activity Context Can Be Misread
The same heart-rate or HRV value can mean different things during pain-limited walking, controlled skating, motorcycle riding, passive passenger transportation, recovery, illness, heat, or emotional stress. A sensor may record the value perfectly while losing the reason it matters.
Default labels such as “walking,” “exercise,” “recreation,” and “transportation” can also hide the actual mobility function. For some disabled people, fewer steps and more rolling movement may represent better function rather than lower engagement.
Step count
Volume without function
Fewer steps may reflect inactivity—or a successful shift to rolling mobility that preserved access with less walking.
HRV / RMSSD
One label, different time scales
Overnight recovery and targeted session HRV answer different questions even when a product interface gives them similar names.
Longitudinal Plus Episodic Evidence
Keep the Time Scales Distinct but Connected
Longitudinal records can show sustained patterns, while episodic sensors can add specificity to a targeted activity or protocol. Combining them responsibly means retaining their different source scopes.
Longitudinal
WHOOP activity and recovery
Repeated activity, strain, sleep, and overnight recovery context support trend review across time.
Functional travel
Strava routes and output
Route, distance, duration, speed, and repeated mobility add real-world functional context.
Episodic
Polar H10 and Kubios
Targeted RRI, HRV/RMSSD, ECG, acceleration, and movement evidence add session-specific detail where available.
WHOOP overnight recovery HRV/RMSSD is not Kubios session HRV. Connection should preserve source identity rather than flattening both into a generic “HRV” feature.
Movement and Transportation Context
Mobility Does Not Stop at the Activity Label
Walking, controlled inline skating, motorcycle control, and passenger transportation couple the body to movement differently. Vehicle, seat, posture, route, duration, vibration, impact, and the person's ability to control or interrupt exposure can all matter.
Movement sequence
Walking and skating use different support and propulsion strategies. A classifier should not infer one from the other.
See the movement contextBody coupling
Active controlled movement and passive passenger exposure can create different profiles even when both appear under a broad mobility category.
Review transportation contextWhy Cohorts Matter
Compare Like with Like Before Looking for Patterns
Within-person baselines and carefully defined cohorts can preserve activity, movement, sensor, transportation, and recovery context. They also make missingness and small samples visible.
Preserve source labels
Keep platform labels and user-confirmed context so classification provenance can be reviewed.
Use within-person baselines
Personal patterns may be more informative than forcing every person into a population-average activity model.
Do not infer membership
Insufficient or ambiguous records should remain review-required rather than silently joining a cohort.
Explainable Evidence Architecture
Provenance and Boundaries Are Product Features
The FSI/CSS Evidence Observatory reconciles source-linked evidence on the server side. A consistent evidence interface and versioned publication contract let public websites display approved values without querying the warehouse or recomputing science.
Fractal Stability Index
FSI answers its named stability question within an established scope. It is not a pain score.
Cohort Similarity Score
CSS answers a similarity question only under an established cohort-comparison scope.
Tensor v0.3
Tensor v0.3 is the frozen multidimensional model identity. It is not another label for FSI or CSS.
Human review remains necessary. An explainable evidence system should show source, units, samples, missingness, limitations, and the path from observation to interpretation.
Opportunities for Wearable Teams
Design Questions Worth Testing
Within-person evidence makes the context problem concrete enough to test rather than leaving it as a generic personalization goal.
Context-preserving classification
Reconcile device labels with user-confirmed mobility function and environment.
Multimodal sensor fusion
Connect physiology, routes, motion, transportation, and documents without erasing source roles.
Personalized mobility intelligence
Use within-person baselines and repeatability rather than treating step volume as the only positive outcome.
Explainable evidence systems
Make provenance, samples, units, missingness, and limitations visible to reviewers.
User-controlled summaries
Explore privacy-preserving ways for people to share relevant functional context with clinicians, accessibility teams, researchers, or agencies.
Timing-aware interpretation
Keep overnight recovery, active-session physiology, and post-activity response distinct.
Opportunities for Mobility and Transportation Teams
Treat the Trip as a Movement Environment
Mobility intelligence can preserve what happens before boarding, how the body couples to a vehicle, what the route imposes, and whether useful function remains after arrival.
Transportation analytics
Preserve vehicle, body-coupling, posture, route, exposure, and shared-ride context.
Alternative-mode comparison
Compare walking, rolling, seated travel, rail, and active-control transport by function and burden rather than convention alone.
Accessible route design
Connect pickup, waiting, transfers, aid handling, ride duration, destination function, and recovery.
Research and Collaboration
A Concrete Case for Responsible Review
This record may support carefully scoped research in wearable context preservation, mobility classification, multimodal evidence, transportation exposure, accessibility, explainability, source provenance, and human-in-the-loop review.
Potential collaboration begins with an explicit research question, privacy scope, validation plan, and agreement about what each source can support.
Review the Evidence Observatory
Move from Product Question to Inspectable Evidence
The public platform page explains how sources are reconciled, how FSI, CSS, and Tensor remain separate, and how publication contracts protect the websites from scientific drift.