Noise
How loud a place is — traffic and busy corridors drive it up.
- SOURCE
- OSM road network (proxy) → ESP32 microphones
- PRECISION
- ~100–300 m modeled · 10–30 m with sensors
- LIMITATION
- Currently inferred from proximity to major roads, not measured.
Four signals.
One city-scale stress layer.
EvoComb blends noise, crowding, heat and air quality into a single Environmental Stress Index across Delhi NCR — street-level where data allows, honest where it does not.
- 4
- signals
- H3 r8
- hex grid
- hourly
- refresh
scroll to reveal ↓
02 · Signals
Four signals, measured separately.
Each layer is scored 0–100 on its own terms before anything is combined. Some are sensor-grade; some are open-data proxies. We say which.
30% · dB
Crowding
How packed an area gets with people and activity.
- SOURCE
- OSM points-of-interest density (proxy) → ESP32 BLE scans
- PRECISION
- ~150 m proxy · 20–50 m with sensors
- LIMITATION
- An activity/footfall proxy — not a live count of people.
Heat
How hot it runs — concrete traps heat, greenery cools it.
- SOURCE
- MET Norway temperature forecast
- PRECISION
- ~km grid, interpolated to cells
- LIMITATION
- Spatially smooth; misses fine-grained heat islands.
Air quality
How clean or polluted the air is to breathe.
- SOURCE
- Open-Meteo air-quality model (US AQI)
- PRECISION
- ~11 km grid, interpolated (IDW) to cells
- LIMITATION
- Modeled and coarse; not a street-level monitor.
03 · Why it matters
These signals carry measurable cost.
Noise
ischaemic heart-disease risk per +10 dB of road-traffic noise
WHO Regional Office for Europe · World Health Organization (2018) ↗Air pollution
deaths in India attributable to air pollution in 2019 (17.8% of all deaths)
Pandey A, Brauer M, Cropper ML, et al. · The Lancet Planetary Health 5(1):e25–e38 (2021) ↗Heat
excess deaths across Europe in the 2003 heatwave
Robine JM, Cheung SLK, Le Roy S, et al. · Comptes Rendus Biologies 331(2):171–178 (2008) ↗Built for planning, research and public-health context — not diagnosis.
04 · The index
One transparent formula. No black box.
The Environmental Stress Index is a published weighted sum — drag a factor to see the cell respond. It is not a trained model, and not human stress.
Noise30%78
Crowding25%64
Heat20%70
Air quality25%82
Greenery−25%20
ESI =30%·Noise+25%·Crowding+20%·Heat+25%·Air quality−greenery
Sample cell · H3
69
Elevated
ESI 0–100
05 · Pipeline
From raw feeds to a hex grid.
Reproducible and open at every stage — the same path every cell takes, every ingest.
- 01
Collect
Open-Meteo AQI, MET Norway heat, OSM roads & POIs — pulled per ingest.
- 02
Normalize
Each raw signal mapped to a comparable 0–100 score.
- 03
Weight
Blended by published weights; greenery subtracts.
- 04
Aggregate
Rolled up to H3 r8 hexagonal cells for the map.
06 · Map
Explore it cell by cell.
Layer
Hover a cell for its readout
Preview is illustrative · the live map serves real ingested cells from /stress/cells.
07 · Evidence
The evidence, organized.
Every health claim above traces to a published source. Open any card for the detail, DOI and how far we trust it.
Every +10 dB of road-traffic noise raises ischaemic heart-disease risk by about 8%.
Environmental Noise Guidelines for the European Region · WHO Regional Office for Europe · 2018
High confidence
The WHO guideline pools long-term cohort evidence linking chronic road-traffic noise to hypertension and ischaemic heart disease, and sets a 53 dB Lden recommendation. Mechanism (stress-hormone and blood-pressure pathways) is detailed in Münzel et al., JACC 2018.
Note · Consistent exposure–response across European cohorts above ~50 dB Lden.
DOI 10.1016/j.jacc.2017.12.015Open source
1.67 million deaths in India were attributable to air pollution in 2019 — 17.8% of all deaths.
Health and economic impact of air pollution in the states of India, GBD 2019 · The Lancet Planetary Health · 2021
High confidence
The Global Burden of Disease 2019 India analysis attributes 1.67M deaths and a measurable GDP loss to ambient and household air pollution, with PM2.5 the dominant driver of cardiopulmonary mortality.
Note · Modelled attributable burden; exposure estimates carry spatial uncertainty.
DOI 10.1016/S2542-5196(20)30298-9Open source
The 2003 European heatwave caused more than 70,000 excess deaths.
Death toll exceeded 70,000 in Europe during the summer of 2003 · Comptes Rendus Biologies · 2008
High confidence
Robine et al. reconstructed national mortality across 16 countries and found >70,000 excess deaths, concentrated in elderly and urban populations where concrete and low greenery intensify heat exposure.
Note · Event total is firm; urban heat-island amplification varies by city fabric.
DOI 10.1016/j.crvi.2007.12.001Open source
WHO sets the safe annual PM2.5 limit at 5 µg/m³. Delhi routinely runs many times higher.
Global Air Quality Guidelines · World Health Organization · 2021
Moderate confidence
The 2021 WHO guideline lowered the annual PM2.5 limit from 10 to 5 µg/m³. Delhi NCR winter averages frequently exceed it by an order of magnitude, which is why EvoComb treats AQI as a heavily weighted signal.
Note · Guideline value is firm; local exceedance depends on station and season.
Coarse, city-scale exposure data hides street-level extremes — so EvoComb states where it is modelled.
Mortality risk attributable to high and low ambient temperature (multicountry study) · The Lancet · 2015
Known limitation
Gasparrini et al. show environmental–mortality risk varies sharply with local exposure, which km-scale grids cannot resolve. EvoComb is street-level only where sensor data exists and labels every modelled cell as such, rather than implying precision it does not have.
Note · Noise and crowd layers are OSM-derived proxies; AQI and heat are ~km-grid models interpolated to cells.
DOI 10.1016/S0140-6736(14)62114-0Open source
08 · Start
See Delhi NCR as one stress layer.
Street-level where data allows. Honest where it does not. Built for planning, research and public-health context.