🚧 The EHSPI live link will be updated shortly. In the meantime, reach us at ops@exposomic.ai
Research & Methodology

A transparency page, not a publications page.
Working in the open.

vasus.ai makes deterministic, evidence-weighted decisions at every layer of the platform — this page documents them. The EHSPI weight vectors are derived from peer-reviewed environmental health literature and currently being submitted for informal methodological review by domain researchers across migraine, respiratory, cardiovascular, sleep, and allergy science. The Google Environmental API Spot Study, comparing our input signals against independent reference networks including OpenAQ and ECMWF, is live and running.

Nothing here is dressed up. We publish the exact weights used in production, confidence tiers on every city score, and document clearly what is validated and what is not. If you are a researcher and see something worth challenging, we want to hear from you.

View EHSPI™ methodology ↓ View EHSPI™ real-time feed ↓ Register for research updates
EHSPI™ v1.6 · Updated nightly

A composite location-scoring system that quantifies relative environmental burden across five chronic condition sensitivities.

The EHSPI scores locations 0–100 where 100 represents the most environmentally favourable conditions in the monitored set and 0 the least. Scores are relative to the 20-city monitoring cohort, not absolute thresholds.

Why a composite index? The exposome is multi-stressor by nature. No single environmental signal adequately characterises health risk. A barometric pressure drop that is irrelevant to a respiratory patient is the primary signal for a migraine patient. The EHSPI aggregates across seven signal types, weighted by condition-specific clinical evidence — producing a score that is meaningfully different for each sensitivity.

How it differs from AQI. Air Quality Index is a single-signal public health measure. The EHSPI is condition-specific: five separate scoring models, each with distinct signal weights derived from published environmental health literature. A city with excellent AQI may score poorly on migraine sensitivity if barometric pressure is highly volatile.

Five-stage computation
1
Environmental signal ingestion
7 Google Environmental API types ingested per tile per refresh cycle. Stored in env_hourly_cache (48h TTL) and pollen_daily_cache (5d TTL).
2
Feature window computation
Raw signals transformed into 24h and 72h window features: pressure_delta_mb, pm25_mean, pollen_exposure_load, heat_index_mean, overnight_hot_hours, and 14 more.
3
Signal normalisation
Each feature normalised across the full monitored tile set using min-max scaling. Ensures scores are relative to the monitored cohort, not absolute.
4
Condition-specific weighted scoring
Normalised features multiplied by condition-specific weights (Σ=1.00 per sensitivity). Five separate scoring vectors — see weight tables below.
5
Composite average
The composite EHSPI™ score is the unweighted average of the five sensitivity scores. Reported alongside individual sensitivity scores.
View EHSPI™ real-time feed ↓
Data maturity & confidence levels

Each city's score includes a confidence label based on how many days of data have been ingested. New cities start at Minimal and mature over 90 days. This honesty about data maturity is a credibility signal, not a weakness — it means scores can be trusted proportionally to their history.

Minimal
< 7 days · < 168 rows
New tile. Score directionally indicative only.
Low
7–29 days · 168–696 rows
Emerging pattern. Use with caution.
Moderate
30–89 days · 720–2,136 rows
Reliable pattern established.
High
90+ days · ~4,320 rows
Full confidence. Seasonality captured.

The EHSPI is not the first index to weight environmental signals by health outcomes. It builds on two established methodologies.

🇨🇦
Canada's Air Quality Health Index (AQHI)
Health Canada & Environment and Climate Change Canada

The AQHI made a foundational shift: instead of reporting raw pollutant concentrations against general thresholds, it combines PM2.5, O3, and NO2 into a single score weighted by their relative contribution to short-term mortality risk. The signal selection is driven by health outcomes, not engineering convenience.

The EHSPI applies identical philosophy — feature selection and weighting grounded in clinical evidence — but extends it to five distinct chronic condition sensitivities and a longitudinal, multi-city framing. Where AQHI asks "what is the risk to the general population right now?", the EHSPI asks "what is the environmental burden on this specific condition, in this city, across the past 72 hours?"

📊
Yale Environmental Performance Index (EPI)
Yale Center for Environmental Law & Policy

The Yale EPI constructs composite environmental quality scores by normalising indicators across countries and applying evidence-based weights — a deductive composite index methodology. It demonstrates that a principled, transparent weighting scheme can produce a credible, comparable cross-location score from heterogeneous data sources.

The EHSPI applies the same deductive composite methodology, but at city level rather than national level, using temporally-windowed derived features rather than annual statistics, and targeting condition-specific health outcomes rather than general environmental quality.

🔬
EXPANSE project & Exposome-NL — urban exposome signal typology
Utrecht University / Institute for Risk Assessment Sciences (IRAS) · EU Horizon 2020

The EXPANSE project (EXposome Powered tools for healthy living in urbAN Settings in Europe) is a major EU Horizon programme led by Exposome-NL at Utrecht University that characterises the urban exposome — the totality of environmental exposures in city environments — and their relationship to chronic disease. EXPANSE defines and validates which environmental signal types are most relevant to urban health burden: air quality, thermal stress, green space, noise, and built environment.

The EHSPI's environmental feature set — specifically the selection of PM2.5, ozone, heat index, pressure volatility, pollen, and AQI duration as the signal categories worth ingesting and weighting — is grounded in the same urban exposome signal typology that EXPANSE has validated at population level across European cities. The EHSPI does not use EXPANSE data directly, but the scientific basis for which signals matter and why draws on the same body of evidence. We are in contact with researchers at Exposome-NL regarding informal methodological review.

How the EHSPI extends both

The EHSPI is best described as: AQHI philosophy (health-outcome-relevant feature selection and weighting) + EPI methodology (deductive composite index, normalised across a monitored cohort) + two novel elements that exist in neither predecessor: condition-specific weight vectors (five separate scoring models, not one), and permanently-stored temporally-windowed features (24h and 72h derived signals, not snapshot readings). No existing index combines all four properties.

🌡️
EHSPI as a climate change proxy

Because the EHSPI is computed nightly and stored permanently, the growing time series is itself a longitudinal record of environmental health burden at city level — and therefore a proxy for how climate change is translating into chronic condition risk over time. Are pollen seasons in Paris extending? Is overnight heat in Dubai becoming structurally worse for sleep? Is pressure volatility in London increasing? The EHSPI time series will surface these shifts at a resolution no annual index can.

The leading framework for tracking climate's health impact is the Lancet Countdown on Health and Climate Change — an annual assessment involving 122 researchers across 57 institutions, tracking 56 indicators across national and regional populations.1 Its 2024 report found that 10 of 15 health-relevant climate indicators reached new records, with people exposed to an average of 50 more days of health-threatening temperatures than would be expected without climate change.2 The Lancet Countdown operates at national and global scale, publishing annually. It tracks what is happening to populations. It does not produce city-level, condition-specific, daily scores.

The EHSPI fills a different layer of the same need: a permanent, nightly record of environmental health burden across 20 cities, disaggregated by chronic condition sensitivity. Over time, it becomes a climate signal at the resolution that population-level annual reporting cannot provide — and one that is directly actionable for the digital health platforms, insurers, and population health programmes that the Lancet Countdown's national-level data cannot serve.

References

1 Watts N, et al. The Lancet Countdown: tracking progress on health and climate change. The Lancet. 2017;389(10074):1151–1164.

2 Romanello M, et al. The 2024 report of the Lancet Countdown on health and climate change: facing record-breaking threats from delayed action. The Lancet. 2024;404(10465):1847–1896.

All five sensitivity weight tables, published transparently.

These are the exact weights used in the live computation. Each signal includes the clinical mechanism in plain language. No other commercial environmental health platform publishes its clinical weighting methodology at this level of detail.

🌀
Migraine & Weather sensitivity
Pressure-led. No pollen neurological mechanism. v1.6 unchanged.
Σ = 1.00
SignalWeightClinical rationale
Barometric pressure delta (24h abs mean)0.38Rapid frontal passage — the strongest replicated environmental trigger for migraine. Chinook/foehn wind literature. Key signal in barometric pressure meta-analyses.
Pressure volatility (std dev over 72h)0.28Sustained variability — distinct from a single delta. Chronic migraineurs show sensitivity to prolonged pressure instability, not just acute drops.
Peak temperature (90th percentile, 6h)0.12Heat as secondary trigger. Summer migraine seasonality. Separate from pressure signal.
Heat index mean (72h)0.10Apparent temperature — integrates humidity and heat. More clinically relevant than temperature alone.
Dew point mean0.06Humidity-headache association. Secondary signal, modest effect size in literature.
AQI peak0.04Air quality as minor trigger. Contextual — included for completeness, low weight.
PM2.5 mean0.02Weak association with migraine. Contextual only.
🫁
Breathing & Air Quality sensitivity
Pollen added as co-primary in v1.6. PM2.5 reduced from 0.28 to 0.22 to accommodate.
Σ = 1.00
SignalWeightClinical rationale
PM2.5 mean (72h)0.22Chronic respiratory trigger. PM2.5-asthma dose-response well-replicated. Reduced from 0.28 in v1.5 to accommodate pollen.
Pollen exposure load (recency-weighted composite)0.18NEW in v1.6. Pollen is the leading precipitant of asthma exacerbations globally. Recency-weighted composite across tree, grass, and weed pollen.
AQI hours unhealthy (72h)0.18Duration of dangerous air quality — chronic exposure metric rather than peak.
Ozone peak (8h rolling mean, 90th percentile)0.16Asthma exacerbation driver. WHO 8h guideline reference. Independent of PM2.5 pathway.
PM2.5 peak (6h, 90th percentile)0.10Acute episode risk. Peak exposure drives exacerbation more than mean in short-onset conditions.
PM2.5 hours above WHO guideline0.08WHO 24h threshold (15 µg/m³) breach frequency.
AQI peak (daily worst)0.04Worst-case AQI reading as a daily sentinel signal.
Temperature mean0.03Cold air bronchospasm — included as secondary contextual signal.
Dew point mean0.01Humidity modulates airway reactivity. Very low weight — minor contextual role.
🔥
Heat & Physical Stress sensitivity
Heat-led. PM2.5 secondary. Weak/indirect pollen-CV link. v1.6 unchanged.
Σ = 1.00
SignalWeightClinical rationale
Heat index mean (72h)0.25Strongest CV heat stress signal. Heatwave mortality literature — sustained heat index drives cardiac strain independently of peak temperature.
Peak temperature (90th percentile)0.20Direct thermal load. Temperature-mortality relationship is approximately linear above 25°C in at-risk populations.
PM2.5 mean0.18PM2.5-cardiovascular mortality link. One of the most replicated relationships in environmental epidemiology. GBD 2019.
AQI hours unhealthy0.14Duration of AQ risk — chronic exposure to elevated AQI has independent CV effects.
Multi-stressor exposure burden0.10Composite of PM2.5, AQI, and temperature. Captures compound exposure effects that single signals miss.
Overnight hot hours (>24°C, 22:00–06:00)0.07Sleep deprivation from thermal discomfort → cardiovascular strain. Obradovich et al.
Temperature mean (72h)0.04Baseline thermal environment. Contextual.
Ozone peak (8h rolling)0.02Ozone-CV association. Weaker than PM2.5 pathway but documented.
🌙
Sleep & Nighttime sensitivity
Overnight heat dominant. No direct pollen-sleep mechanism. v1.6 unchanged.
Σ = 1.00
SignalWeightClinical rationale
Overnight hot hours (22:00–06:00, >24°C)0.35Primary sleep disruption driver. Obradovich et al. 2017: 1°C increase in minimum temperature → 0.58% more insufficient sleep nights.
Heat index mean (72h)0.25Humid heat disrupts sleep more than dry heat. 72h window captures multi-night heat event persistence.
Pressure volatility (std dev)0.14Pressure changes affect sleep architecture. Barometric variability correlated with poor sleep quality in sensitive populations.
Dew point mean0.12Dew point >21°C is a critical thermal comfort threshold. Oppressive humidity independently disrupts sleep.
AQI peak0.08Air quality affects sleep quality. Upper airway irritation and respiratory arousal from pollution.
PM2.5 mean0.06Chronic PM2.5 exposure linked to sleep-disordered breathing and OSA exacerbation.
🌸
Allergies & Pollen sensitivity
NEW in v1.6. Pollen-led. Highest single weight (0.40) across all five sensitivities.
Σ = 1.00
SignalWeightClinical rationale
Pollen exposure load (recency-weighted composite)0.40The dominant driver — the highest single weight across all five sensitivities. D'Amato et al. 2020: outdoor pollen is the leading cause of allergic rhinitis globally. Recency-weighted to prioritise recent days.
Tree pollen UPI (Universal Pollen Index, peak)0.15Tree pollen (oak, birch, cedar) — spring peak season. Species-specific sensitisation is clinically significant.
Grass pollen UPI (peak)0.12Grass pollen (timothy, bermuda, ryegrass) — summer peak. The most globally prevalent pollen allergen type.
Weed / ragweed pollen UPI (peak)0.10Weed and ragweed — autumn. Ragweed is the single largest allergic trigger in North America and increasingly in Europe.
PM2.5 mean0.10Pollen-PM2.5 compound effect is documented — PM2.5 particles carry pollen fragments and amplify allergenicity.
Ozone peak (8h rolling)0.08Ozone-allergen synergy. Ozone increases nasal epithelial permeability, amplifying pollen sensitisation response.
Dew point mean0.05Humidity modulates pollen dispersal and bursting. High humidity causes pollen grains to rupture, releasing smaller allergenic particles.
📋
Methodology statement

The EHSPI weighting methodology is evidence-informed, derived from published environmental health literature, and is currently being evaluated against external reference datasets. A formal validation study is in preparation. We welcome methodological review from research institutions.

Register for research updates →

20 cities. Global coverage. Three priority tiers.

Each city is assigned a priority tier that determines its refresh interval. Highest-priority tiles refresh every 60 minutes; standard tiles every 180 minutes. Data confidence reaches High after approximately 90 days of continuous ingestion.

City Region Priority Refresh Confidence (at 90d)
Delhi NCR, India South Asia Highest 60 min High
Bangkok, Thailand SE Asia Highest 60 min High
Jakarta, Indonesia SE Asia Highest 60 min High
Los Angeles, USA North America Highest 60 min High
Mexico City, Mexico North America Highest 60 min High
London, UK Europe High 120 min High
Paris, France Europe High 120 min High
Istanbul, Turkey Europe / MENA High 120 min High
Dubai, UAE Middle East High 120 min High
Riyadh, Saudi Arabia Middle East High 120 min High
Singapore SE Asia High 120 min High
Kuala Lumpur, Malaysia SE Asia High 120 min High
Hong Kong East Asia High 120 min High
New York City, USA North America High 120 min High
Colombo, Sri Lanka South Asia High 120 min High
Zurich, Switzerland Europe Standard 180 min Moderate
Mecca, Saudi Arabia Middle East Standard 180 min Moderate
Sydney, Australia Southern Hem. Standard 180 min Moderate
Nairobi, Kenya Africa Standard 180 min Moderate
Cape Town, South Africa Africa Standard 180 min Moderate

Current EHSPI™ scores across all 20 cities.

Sourced from GET /v1/ehspi, updated nightly. Scores are relative to the monitored cohort — 100 = most environmentally favourable in the set.

Loading live EHSPI scores…

Scores refresh on page load · Computed nightly · Confidence labels reflect data history

Study in progress — findings to be published Q2 2026 (indicative)

Validating our data sources against independent reference networks.

vasus.ai ingests environmental data from three Google APIs: the Air Quality API, the Solar and Weather API, and the Pollen API. As part of our commitment to data transparency, we are conducting a systematic spot study comparing Google Environmental API outputs against independent reference data sources for a representative subset of the 20 monitored cities.

This study matters for three reasons: it validates that the input signals feeding the EHSPI are accurate relative to authoritative monitoring networks; it documents any systematic biases transparently; and it establishes vasus.ai as a platform committed to methodological transparency rather than treating data sources as a black box.

When complete: findings sections populate, data download link appears (raw comparison dataset as CSV, no login required), and the page is promoted more prominently in navigation.

Planned reference sources
OpenAQ
PM2.5, PM10, O₃, NO₂
Global — 10,000+ stations
api.openaq.org
Open-Meteo
Temperature, pressure, humidity
Global — ECMWF gridded
api.open-meteo.com
EEA Air Quality
Hourly monitored AQ
EU cities incl. London, Paris, Istanbul
EEA portal
SILAM pollen
Pollen forecast model
European cities — FMI/ENII model
silam.fmi.fi
ECMWF Copernicus CAMS
AQ reanalysis — PM2.5, O₃, pollen
Global gold standard
atmosphere.copernicus.eu

EHSPI Research Updates

The EHSPI is a living index. Scores update nightly as new environmental data is ingested across 22 global cities, and the methodology evolves as formal validation progresses. We publish periodic research notes covering: weight vector updates and rationale, validation findings as they emerge, new city additions, and methodological notes on the evidence architecture. If you are a researcher, clinician, or institution working at the intersection of environmental health and chronic conditions, register below to receive these updates.

Your details are used only to send EHSPI research updates. We do not share your information with third parties. You can unsubscribe at any time.