Portfolio Project · Interaction Design · 2022

SOMA
SweatSense

A wearable + app system that translates sweat biomarkers into plain-language guidance — before, during, and after exercise.

Role
Individual project
Type
Wearable + Mobile app
Focus
Biosensor UX / Health tech
S
09:42 · Ready to train? 74 Readiness Na⁺ OK LAC 2.1 HRV 58 Start session →

01 — Problem reframe

The real gap isn't
a missing product.
It's body blindness.

Athletes and casual exercisers alike make every training decision — how hard to push, when to stop, how to recover — based entirely on subjective feeling. There is no real-time data informing these choices.

Existing wearables measure physiological signals: heart rate, steps, SpO₂. What they miss is the biochemical layer — what is actually happening inside the body at a metabolic level during exertion.

Sweat encodes that layer. The question SOMA asks: can we translate it into guidance a non-expert can act on?

85%
Self-treat or ignore pain entirely Questionnaire, n=100. Only 15% see a doctor after sports-related pain or injury.
41%
Report injuries from exercise Most common causes: overloaded intensity, insufficient warm-up, poor body condition awareness.
0
Consumer products detect sweat biomarkers in real time As of 2022, no consumer wearable monitors sweat lactate or electrolytes during exercise.

Active but uninformed. The missing link isn't motivation or access to equipment — it's real-time feedback from the body itself.

— Core design problem, synthesised from user research + market analysis

02 — Science anchor

Why sweat?
What the research says.

Sweat is one of the most studied biofluids in wearable sensor research. It is non-invasive, continuously available during exercise, and contains multiple signals that correlate with metabolic state.

C₃H₆O₃
Lactate
Produced during anaerobic metabolism. Rising sweat lactate signals that muscles are approaching their aerobic threshold — the clearest biochemical indicator of overload during exercise.
Fatigue signal
Na⁺ Cl⁻
Electrolytes
Sodium and chloride concentration reflects hydration status and thermoregulation. Loss above threshold (>2% body mass) measurably impairs physical and cognitive performance.
Hydration risk
C₂₁H₃₀O₅
Cortisol
The primary stress hormone. Elevated sweat cortisol correlates with psychological and physiological overreaching. Wearable electrochemical sensors can detect it in real time in sweat.
Stress / recovery
HRV
Heart Rate Variability
RMSSD, measured at rest in the morning, is validated as a non-invasive proxy for autonomic recovery state. Strong correlation (r=0.84–0.85) with lactate threshold in systematic meta-analyses.
Pre-exercise readiness

Honest design constraint: Sweat lactate as a proxy for blood lactate remains scientifically debated — correlations vary by sweat rate, body site, and individual. Cortisol detection is promising but consumer validation is limited. SOMA's framing is intentionally careful: it presents signals as guidance inputs, not clinical diagnostics. This distinction matters and should be visible in the product's voice.

03 — Product logic

Two moments of decision.
Two different signals.

The product logic maps to two distinct physiological questions that athletes face but currently can't answer with data.

Before exercise

"Is my body ready to train today?"

  • HRV (RMSSD), measured on waking — reflects autonomic nervous system recovery. Low HRV = incomplete recovery from previous load.
  • Resting skin conductance — baseline stress proxy before session begins.
  • Prior session data — lactate and electrolyte loss from last session informs recovery estimate.
Go Modify intensity Rest
During exercise

"Should I continue, reduce, or stop?"

  • Sweat lactate rate-of-change — rapid rise signals approaching anaerobic threshold, precursor to overload and muscle damage.
  • Na⁺ / Cl⁻ concentration — electrolyte loss tracking; alert when dehydration risk crosses threshold.
  • Sweat rate + temperature — context for electrolyte readings; used to calibrate signal accuracy.
Continue Reduce intensity Stop + recover

04 — User research

Quantitative survey + in-depth interviews.

n=100 questionnaire to map injury patterns, plus three in-depth interviews across a spectrum of exercise engagement — from sporadic beginner to data-driven enthusiast.

Lisa, 22
Student · Yoga & evening jogs · Sporadic
Beginner

"I exercise freely but have no regularity — hard to stay consistent."

Relies entirely on feeling to gauge effort. No awareness of concepts like warm-up or recovery. When pain occurs, default response is to ignore it or stop without understanding why. Needs guidance that doesn't require prior sports knowledge.

Ada, 30
Employer · Running & gym · 5× week
Enthusiast

"I need data to tell me what my body actually needs."

Trains consistently and tracks data, but has hit a plateau. Current devices give heart rate and steps — not the metabolic insight she needs to break through. Frustrated by conflicting advice online. Wants expert-level feedback without a sports scientist.

Lily Xiong, 27 — Student · Yoga + runs · ~5×/week

"Pain occurred during yoga — I just ignored it."

Classic body blindness pattern: discomfort without context leads to dismissal. No feedback loop between pain signal and behaviour change.

Nancy Yang, 31 — Business Owner · Outdoor sports · Uses sports apps

"Good habits = zero injuries so far."

Structured warm-up and cool-down replaces the need for real-time data. Demonstrates that behaviour, not just technology, drives injury prevention — SOMA's design must reinforce habit formation, not just surface numbers.

Steve Qin, 30 — Postdoc · Football · Fri & Sun with teammates

"Recovery took very long — still have sequelae."

Insufficient warm-up led to ankle and knee injuries with lasting effects. Represents the highest-stakes case: overload that wasn't detected until after the injury occurred. Exactly the scenario SOMA's in-session alert is designed to prevent.

05 — User journey

Where SOMA intervenes
across the exercise lifecycle.

Before exercise
  • Search tips and training plans
  • Attempt warm-up routines
  • Assess own readiness by feel
Pain point: No objective readiness signal. Users proceed regardless of actual recovery state.
SOMA: HRV-based readiness score + warm-up intensity recommendation on session start.
During exercise
  • Exercise per memorised instructions
  • Self-monitor fatigue by subjective feel
  • Cool down (frequently skipped)
Pain point: No real-time overload detection. Users push past safe limits without knowing.
SOMA: Live sweat lactate + Na⁺ monitoring. Alert modal when overload threshold is exceeded.
After exercise
  • Log data manually or not at all
  • Self-assess session effectiveness
  • Address pain by ignoring or self-treating
Pain point: No data-driven recovery guidance. Decisions about next session are uninformed.
SOMA: Session summary with sweat data, hydration loss, fatigue index, and AI recovery recommendation.
06 — Design principles

Three principles that shaped every decision.

P.01

Plain language, not raw data

Sweat biomarkers are meaningless to non-experts. Every sensor reading must be translated into an action the user can take. "Your lactate is rising fast" becomes "Reduce your pace."

P.02

Guidance, not surveillance

The system earns its place by being useful at critical moments, not by collecting data continuously. Alerts should feel like a knowledgeable training partner — not an alarm system.

P.03

Honest uncertainty

The science of sweat biomarkers is still maturing. SOMA's UI language reflects this — signals are framed as inputs for decision-making, not clinical facts. This trust boundary is a design choice, not a disclaimer.


07 — Key screens

Six screens across
the exercise journey.

Each screen owns a specific moment in the Before / During / After structure. No screen does more than one job.

09:42
Good morning
Body readiness
Good
HRV 58 · Well rested
Na⁺ level
Normal
Last session
32 min
Lactate
2.1 mM
Recovery
Full
Start session →
Before
Dashboard
Pre-session readiness with HRV-based body state summary
Live
42% sweat index
LAC
3.8
Na⁺
OK
HR
142
Fatigue indexLow
Hydration68%
During
Live monitor
Real-time sweat ring, fatigue index and electrolyte status
Live
⚠️
Overload detected
Lactate rising fast. Your body is approaching its limit. Reduce pace now to prevent injury.
Reduce pace
Finish
During — Alert
Overload alert
Modal triggered when lactate threshold is exceeded
10:18
Session complete
32 min · Moderate intensity
Sweat loss
12 ml
Peak lactate
5.2 mM
Na⁺ loss
Low
Fatigue index
68%
AI recovery tip
Drink 400ml water in the next 30 min. Light stretching recommended. Next hard session: 48h rest advised.
After
Session complete
Post-session sweat data summary with AI recovery recommendation
History
This week
MTWTFSS
Thu
32 min · 5.2 mM peak
Good
Tue
45 min · 7.1 mM peak
High
Mon
28 min · 3.8 mM peak
Good
After
History
Weekly trend with session log and peak lactate per session
SOMA OK body ready microfluidic sensor patch sweat sensor array
Hardware
SOMA band
Upper-arm wearable with microfluidic biosensor array and wireless app sync

08 — Competitive positioning

Not a better smartwatch.
A different instrument.

The key insight from competitive analysis: Apple Watch and fitness trackers measure physiological signals. SOMA measures biochemical ones. These are fundamentally different layers of information — not competing features.

Product HR / HRV Sweat lactate Electrolytes Cortisol Real-time guidance Plain language output
Apple Watch Ultra Partial Partial
Garmin Fenix Partial
Whoop 4.0
Smart clothing (Hexoskin)
SOMA SweatSense

09 — Project outcomes

What SOMA delivers.

01

Sweat-sensing wearable

Upper-arm band with microfluidic biosensor array detecting sweat lactate, sodium, cortisol, and pH in real time. Designed for wearability during high-intensity exercise.

02

Companion app system

Plain-language guidance across Before / During / After — readiness score, live overload alerts, and a data-driven recovery plan. Science translated into decisions a non-expert can make.

03

First design of its kind

Addresses sweat biomarker detection and fatigue prevention together in a single consumer product — bridging the gap between sports science research and accessible health technology.

10 — Reflection

What worked.
What I'd do differently.

What worked

  • Grounding the product concept in peer-reviewed science gave the design rationale credibility beyond typical student projects.
  • The Before / During / After structure prevented feature creep — each screen has exactly one job.
  • Acknowledging scientific uncertainty in the UI language (signals as guidance, not facts) strengthened rather than weakened trust.
  • User research spanning both beginner and enthusiast profiles revealed a shared blind spot: both personas lack real-time body awareness, just for different reasons.

If I continued this project

  • Validate sweat lactate correlation on a broader population sample — the current science is promising but contested.
  • Prototype the alert UX with real athletes to calibrate false-positive tolerance. Too many alerts = ignored alerts.
  • Design a personalised calibration flow — biomarker baselines vary significantly between individuals, and the app needs to account for this before making reliable recommendations.
  • Explore partnership with sports science institutions for clinical validation, which would be necessary before any health claims could be made in a real product.