Luminate Research
A longitudinal environmental health platform
Daily lives. Population evidence.
Luminate uses the phone people already carry to organise mobility, activity, place and modelled exposure across weeks, months and potentially years. Its app architecture and exposure calculations are built on patent-pending technology designed for population-scale deployment.
Patent-pending technologyDesigned for scaleParticipant-correctable
01 / THE OPPORTUNITY
Scale and duration change the questions we can ask.
A low-friction app allows for research campaigns at a breadth and duration that short studies struggle to capture: repeated commutes, seasonal shifts, life events and, crucially, whether new behaviours last.
COHORT VIEW · 1,248 PARTICIPANT-DAYSResearch patterns emerging across the city
08:10 PEAKBus · 7.8 km34 min · NO₂ 38
18:05 PEAKUnderground · 9.1 km31 min · NO₂ 53
ACTIVE ROUTEWalk + rail · 6.4 km44 min · NO₂ 19
LOWERHIGHER APP EXPOSURE
MODELLED EXPOSURE · CORRIDOR
NO₂ 38O₃ 44PM₁₀ 29PM₂.₅ 17
3 recurring corridors
MONTH 1Road corridor31.4 NO₂ · 42 active min/week
MONTH 6Greenway + rail18.6 NO₂ · 76 active min/week
ACTIVE TRAVEL
+34 min/week, sustained
COMMUTESEXPOSURECHANGE
A role for sensors
Personal sensors remain valuable for direct measurement and validation. Luminate can complement them: sensor substudies can quantify model performance, while the app extends mobility and exposure context across larger cohorts and longer periods. The research proposition is not app versus sensor—it is using each where it is strongest.
02 / THE RESEARCH DATA PIPELINE
From raw phone signals to interpretable research data.
Luminate turns noisy, asynchronous observations into a reviewable daily record, then into exposure, mobility and behaviour summaries that retain their connection to the underlying episode.
Location + motion
Adaptive background sampling captures movement, device motion and time while balancing continuity against battery cost.
Journeys + stops
Noise-handled trajectories become bounded journeys and stationary visits rather than an undifferentiated stream of coordinates.
Mode + place
Classifiers and recurring-place histories add meaning, while confidence and participant corrections remain explicit.
Exposure + dose
Timed routes, stops and activity modes are matched to pollution fields for each ended, participant-confirmed interval.
Patterns + questions
Daily records become comparable periods, intervention outcomes and traceable natural-language answers.
EXPOSURE / JOURNEYS + STATIONARY TIME
A complete day, not routes alone.
Confirmed journeys use timed route coordinates and activity mode. Stops are also modelled, so home, work and other stationary periods contribute to the same complete-day account.
- NO₂, O₃, PM₁₀ and PM₂.₅
- Average concentration and modelled dose
- Time-weighted summaries with explicit coverage
FRIDAY · COMPLETE DAYJourneys + stops
OVERALL EXPOSUREAverage
NO₂ 21O₃ 48PM₁₀ 18PM₂.₅ 9
At home ✓8h 10m
NO₂ 14O₃ 43PM₁₀ 16PM₂.₅ 8
Walking ✓22m
NO₂ 18O₃ 47PM₁₀ 17PM₂.₅ 9
Underground ✓31m
NO₂ 53O₃ 18PM₁₀ 41PM₂.₅ 24
Office ✓8h 13m
NO₂ 21O₃ 48PM₁₀ 18PM₂.₅ 11
Cycling ✓27m
NO₂ 13O₃ 52PM₁₀ 14PM₂.₅ 7
Evening stop ✓6h 13m
NO₂ 17O₃ 49PM₁₀ 16PM₂.₅ 9
03 / THE PARTICIPANT EXPERIENCE
A useful app for participants. Infrastructure for researchers.
Participants receive a legible timeline, exposure feedback and useful personal insights. Their review actions improve research provenance without turning every day into a questionnaire.
01Review the day
9:41Recording quietly
Friday, 18 July
OVERALL EXPOSUREAverage
NO₂ 21O₃ 48PM₁₀ 18PM₂.₅ 9
Start of day1h 28m
Walking · 91%22m
Bus · 87%17m
Office7h 46m
02See the pattern
9:4130 days tracked
Insights
AverageDose
NO₂ 23O₃ 49PM₁₀ 21PM₂.₅ 12
−9%vs previous 30 days
Journeys + places
WWalking10.2
CCycling13.6
UUnderground38.4
POffice22.6
03Ask why
9:41Ask Luminate
Ask Luminate
When was my NO₂ exposure highest last week?
HIGHEST NO₂ JOURNEY
Tuesday · Underground
52.7 µg/m³94% model coverage
Σ(concentration × duration) ÷ 21 min
Ask a follow-up… ↑
VISIBLE TO PARTICIPANTS
Suggestions, confidence, exposure and coverage remain legible and correctable.
USEFUL TO RESEARCHERSEach answer stays connected to the underlying episode and participant action.
04 / LONGITUDINAL OUTCOMES
Measure the pattern before, during and after an intervention.
A shared timeline creates comparable outcomes across time: exposure, travel mode, active minutes, route choice, timing and time spent at places.
Baseline travel patterns
Average modelled NO₂ exposure and observed travel across the study cohort.
Average NO₂ concentration µg/m³
Walking18 µg/m³286 people · 412h · 1,086 mi
Cycling15 µg/m³142 people · 226h · 2,018 mi
Bus28 µg/m³394 people · 731h · 6,442 mi
Underground / Tube38 µg/m³318 people · 604h · 8,190 mi
Running16 µg/m³68 people · 74h · 438 mi
Car32 µg/m³451 people · 910h · 12,840 mi
→Research intervention
Changed travel patterns
The same measures reveal substantial mode substitution and lower-exposure choices after intervention.
Average NO₂ concentration µg/m³
Walking11 µg/m³364 people · 588h · 1,574 mi
Cycling9 µg/m³221 people · 389h · 3,612 mi
Bus19 µg/m³351 people · 602h · 5,201 mi
Underground / Tube23 µg/m³277 people · 486h · 6,771 mi
Running10 µg/m³103 people · 126h · 764 mi
Car21 µg/m³329 people · 641h · 8,906 mi
THE QUESTIONS
Luminate helps answer the research questions.
Where and when does exposure occur?
Decompose modelled exposure by hour, place, journey, mode and day type while retaining the episode behind each aggregate.
How do routines change over time?
Characterise mode share, route choice, active travel and recurring places across seasons, life events and interventions.
Does personalised information change behaviour?
Test whether feedback precedes cleaner routes, mode substitution, changed timing or increased physical activity—and whether it lasts.
Which intervention works for whom?
Randomise message framing, insight timing or suggested actions, then compare pre-specified outcomes.
What explains the observed pattern?
Trigger brief surveys after relevant journeys or insights to capture constraints, perceptions, reasons and wellbeing.
Where does the model need strengthening?
Use personal-monitor substudies to quantify bias across modes, microenvironments, geographies, devices and groups.
05 / AREAS FOR RESEARCH EXPANSION
Luminate is a research platform — we are keen to expand into new areas of environmental health research.
These are research directions—not current production claims—and would require protocol design, validation, participant consent and appropriate permissions.
Watch, heart rate and sleep context.
Link consented wearable measures to exposure episodes, activity, travel context and accumulated dose.
Noise, pollen, heat and weather.
Attach additional environmental surfaces or observations to the same timed mobility record.
Surveys, A/B tests and adaptive feedback.
Deliver questions or messages in context with transparent assignment and pre-specified outcomes.
06 / SCIENTIFIC FOUNDATIONS
Built using ERG science from the ground up.
Luminate brings four disciplines together in one participant experience—so the technology, the exposure model and the research question are designed as one system.
ENVIRONMENTAL RESEARCH GROUP
Science × people × technology
Multidisciplinary expertise shapes what Luminate measures, how it interprets a day and how people take part.
Understanding the pollutant field.
Air quality measurement, modelling and interpretation ground the app in the science of where pollution comes from and how concentrations vary through space and time.
Connecting pollution to a lived day.
Time, place, route, activity and breathing rate become an interpretable exposure account that includes stationary periods as well as journeys.
Learning what information changes.
Participant motivations, inequalities, trust and behaviour-change methods guide how studies ask questions and how outcomes are understood.
Building with people, not around them.
Mobile engineering, co-design and community participation support a useful app, responsible data collection and research people can sustain over time.
Collaborate with us
Bring a question. Help shape the research.
We welcome collaborations across exposure science, mobility, physical activity, epidemiology, digital interventions, participant experience and responsible health-data research.