ENVIRONMENTAL OPERATIONS
City intelligence map
DEMO DATA2026Change password
Live map — measured readings

One map, six readings of the city.

A boundary, a measurement, a facility, an open investigation and a belief, read against each other on one geography.

Detail

Open →
ACTIONABLE INTELLIGENCE

What should the room do first?

Ranked priorities from the figures on this page, each with an owning department and what would prove it wrong. Advice to an officer, not a decision.

Modelled data — no ground stations connectedAll 28 current readings come from the Open-Meteo CAMS model, not physical sensors. CAMS resolves to roughly a 40 km grid, so adjacent zones can return identical figures. Set CPCB_API_KEY to add official ground stations.
CITIZEN ACTION NETWORK

Neighbourhood response pulse

2 reports received · 2 still open · 0 overdue · 0 verified RWAs

Open action queueCitizen report portal ↗

Priority investigations

Signals requiring field verification

All investigations →
Elevated PM2.5 at Sector 37Sector 37 · PM2.5 · confidence LowSector 37 (CAMS modelled) recorded CPCB AQI 312 (Very Poor), PM2.5 88.7 and PM10 121.5 µg/m³ at 2026-11-10 11:00. Modelled value from Open-Meteo CAMS, not a ground station. CAMS resolves to roughly a 40 km grid, so neighbouring zones can return the same figure and a local hotspot is invisible to it. Treat it as a prompt to measure, not as a measurement. Fine-fraction dominated (PM2.5 is 73% of PM10). That pattern is typical of combustion — waste or biomass burning, vehicle exhaust, DG sets. Look there first. First step: Field walk focused on burning, idling and DG sets; suggested owner MCG Sanitation / HSPCB. Confirms or rules out: A co-located ground reading plus geo-tagged field photographs. Rule out regional transport by comparing neighbouring zones over the same hours. Drafted from an Azure OpenAI briefing over this database. Verify on site before any enforcement; no source is attributed.
High

Department action tracker

Operational accountability

All actions →
Ward 17: Brief RWAs and colony representatives across the wardMCG ward team / RWA · Vikas SadanOwner: Ward owner not assigned · Due: 2026-11-11
Open
Ward 17: Check DG-set compliance at large premises in the wardHSPCB / facility operator · Vikas SadanOwner: Ward owner not assigned · Due: 2026-11-11
Open
Ward 17: Sweep and sprinkle the ward's arterial stretches; log coverageMCG / GMDA · Vikas SadanOwner: Ward owner not assigned · Due: 2026-11-11
Open
Ward 17: Notify every school in the ward and confirm receiptDEO / school management · Vikas SadanOwner: Ward owner not assigned · Due: 2026-11-11
Open
Ward 17: Inspect C&D sites and known burning spots in the ward; issue stop-work on breachMCG / HSPCB · Vikas SadanOwner: Ward owner not assigned · Due: 2026-11-11
Open
Manesar: Ban construction and demolition except essential/public projects and specified exemptionsMCG / GMDA · ManesarOwner: Agency lead not assigned · Due: 2026-11-11
Open
Manesar: Close brick kilns, hot-mix plants and stone crushers not on cleaner fuelHSPCB · ManesarOwner: Agency lead not assigned · Due: 2026-11-11
Open
Udyog Vihar: Targeted action at identified air-pollution hotspotsMCG / HSPCB · Udyog ViharOwner: Agency lead not assigned · Due: 2026-11-11
Open
Turn a signal into a verified outcome

Log a reading on the air view, open an investigation on the operations view, let the City Brain propose the action, then record inspection evidence.

Open the City Brain →

Monitoring data.

Capture pollutant measurements with provenance and quality status.

↓ Export CSV

Add a reading

AQI is the CPCB National AQI (CPCB 2014 method): a sub-index per pollutant, the worst one wins, using CPCB's own averaging periods — 24 h for PM/NO2/SO2/NH3, 8 h for CO/O3. It is the index GRAP stages are defined against, so it is not the US AQI some providers return. A dash means CPCB's three-pollutant rule (PM mandatory) was not met — the concentrations are still real. The reading history is in the table above; PM columns are the current hour.

Official geography.

MCG wards and colonies as published by GMDA OneMap, with the operational owner for each.

Manage wards & RWAs →
Administrative layerA ward boundary assigns responsibility; it is not a pollution measurement and carries no air-quality meaning on its own. Colony approval status comes from the MCG record, not from an assessment made here.

City facilities.

Places a field team may have to reach, verify or protect.

+ Add a facility
A marker is a place to verifyA mapped facility is not a finding and not an emission source. Sensitive receptors (schools, hospitals) are mapped so exposure can be considered, not to imply that anything nearby is responsible.

Investigate the signal.

Record anomalies, hypotheses, evidence and confidence for field verification.

Field inspections →
Evidence standardA hotspot is an investigation lead, not proof that a particular source caused pollution. Do not use this record alone for enforcement.

Log a hotspot

Investigation register

All recorded signals · 1

Elevated PM2.5 at Sector 37Sector 37 · PM2.5 · 2026-11-10 17:31

Sector 37 (CAMS modelled) recorded CPCB AQI 312 (Very Poor), PM2.5 88.7 and PM10 121.5 µg/m³ at 2026-11-10 11:00. Modelled value from Open-Meteo CAMS, not a ground station. CAMS resolves to roughly a 40 km grid, so neighbouring zones can return the same figure and a local hotspot is invisible to it. Treat it as a prompt to measure, not as a measurement. Fine-fraction dominated (PM2.5 is 73% of PM10). That pattern is typical of combustion — waste or biomass burning, vehicle exhaust, DG sets. Look there first. First step: Field walk focused on burning, idling and DG sets; suggested owner MCG Sanitation / HSPCB. Confirms or rules out: A co-located ground reading plus geo-tagged field photographs. Rule out regional transport by comparing neighbouring zones over the same hours. Drafted from an Azure OpenAI briefing over this database. Verify on site before any enforcement; no source is attributed.

Confidence: Low · Status: Open · Owner: Unassigned
High

Gurugram Clean Air Brain

A living operational model that connects evidence, beliefs, decisions, interventions and measured outcomes.

How the brain worksIt gathers observations, assesses confidence, creates time-stamped hypotheses, proposes field actions and learns from recorded outcomes. It does not infer legal guilt, claim causation, or execute actions without a human decision.

Latest reasoning cycle

What the system observed and inferred

Cycle #102026-11-10T11:46:58+00:00

Scanned 7 zones; captured 534 evidence records; created 7 temporal zone beliefs and 0 proposals for human review.

7 zones534 evidence items7 beliefs0 proposals0 queued changes
Engine: clean-air-brain/1.0 · Trigger: user:admin

Cognitive operating loop

From raw signals to accountable action

01 · ObserveReadings, citizen reports, source evidence, inspections
02 · ContextualisePlace observations in zone, time and operational context
03 · Form beliefsConfidence, uncertainty, provenance and competing signals
04 · RecommendPrioritised proposals with reasons and evidence links
05 · Govern actionHuman approval, named owner, deadline and work order
06 · LearnRecord outcomes; promote reviewed lessons to operational memory

City world model · Current beliefs

These are provisional hypotheses, not established facts. Open the zone API for evidence lineage.

Udyog ViharActive

Udyog Vihar should be prioritised for clean-air investigation

Confidence46%Uncertainty 55%
Prioritisation score is a heuristic, not an AQI calculation or legal finding. Signals: latest recorded AQI 312; 14-day recorded PM2.5 mean 101.6 µg/m³; 1 citizen reports in 30 days Citizen reports and unverified/demo readings are not treated as verified facts.
34 linked evidence items · Revised 2026-11-10T11:46:58+00:00 Inspect zone evidence ↗
Vikas SadanActive

Vikas Sadan should be prioritised for clean-air investigation

Confidence46%Uncertainty 55%
Prioritisation score is a heuristic, not an AQI calculation or legal finding. Signals: latest recorded AQI 312; 14-day recorded PM2.5 mean 101.6 µg/m³; 1 citizen reports in 30 days Citizen reports and unverified/demo readings are not treated as verified facts.
68 linked evidence items · Revised 2026-11-10T11:46:58+00:00 Inspect zone evidence ↗
Golf Course ExtensionActive

Golf Course Extension should be prioritised for clean-air investigation

Confidence40%Uncertainty 60%
Prioritisation score is a heuristic, not an AQI calculation or legal finding. Signals: latest recorded AQI 312; 14-day recorded PM2.5 mean 101.6 µg/m³ Citizen reports and unverified/demo readings are not treated as verified facts.
153 linked evidence items · Revised 2026-11-10T11:46:58+00:00 Inspect zone evidence ↗
ManesarActive

Manesar should be prioritised for clean-air investigation

Confidence40%Uncertainty 60%
Prioritisation score is a heuristic, not an AQI calculation or legal finding. Signals: latest recorded AQI 315; 14-day recorded PM2.5 mean 97.3 µg/m³ Citizen reports and unverified/demo readings are not treated as verified facts.
67 linked evidence items · Revised 2026-11-10T11:46:58+00:00 Inspect zone evidence ↗
NH-48Active

NH-48 should be prioritised for clean-air investigation

Confidence40%Uncertainty 60%
Prioritisation score is a heuristic, not an AQI calculation or legal finding. Signals: latest recorded AQI 312; 14-day recorded PM2.5 mean 101.6 µg/m³ Citizen reports and unverified/demo readings are not treated as verified facts.
89 linked evidence items · Revised 2026-11-10T11:46:58+00:00 Inspect zone evidence ↗
Sector 37Active

Sector 37 should be prioritised for clean-air investigation

Confidence40%Uncertainty 60%
Prioritisation score is a heuristic, not an AQI calculation or legal finding. Signals: latest recorded AQI 312; 14-day recorded PM2.5 mean 101.6 µg/m³ Citizen reports and unverified/demo readings are not treated as verified facts.
107 linked evidence items · Revised 2026-11-10T11:46:58+00:00 Inspect zone evidence ↗
TERI GramActive

TERI Gram should be prioritised for clean-air investigation

Confidence40%Uncertainty 60%
Prioritisation score is a heuristic, not an AQI calculation or legal finding. Signals: latest recorded AQI 312; 14-day recorded PM2.5 mean 101.6 µg/m³ Citizen reports and unverified/demo readings are not treated as verified facts.
16 linked evidence items · Revised 2026-11-10T11:46:58+00:00 Inspect zone evidence ↗

Decision queue

Brain-generated proposals. Nothing becomes a work order without a named human decision.

9awaiting decision
3distinct recommendations
0already decided

Grouped by recommendation. Each zone still needs its own decision — the evidence behind it differs.

Inspect DG set operating and emission controls

1 zone
Vikas Sadan HSPCB / competent authority
High

Most frequent citizen-observed category in the last 30 days: DG set emissions / noise (1 reports). Validate on site before enforcement.

40% confidence · 65 evidence items
Decide

Conduct a geo-tagged burn-site inspection and repeat check

1 zone
Udyog Vihar MCG Sanitation
High

Most frequent citizen-observed category in the last 30 days: Open garbage burning (1 reports). Validate on site before enforcement.

40% confidence · 31 evidence items
Decide

Run a coordinated hotspot source-verification visit

this signal prioritises investigation but does not identify a source.

7 zones
Udyog Vihar MCG / relevant source-owning authority
High

Latest recorded AQI is 207

46% confidence · 33 evidence items
Decide
Sector 37 MCG / relevant source-owning authority
High

Latest recorded AQI is 207

46% confidence · 107 evidence items
Decide
Golf Course Extension MCG / relevant source-owning authority
High

Latest recorded AQI is 207

46% confidence · 153 evidence items
Decide
NH-48 MCG / relevant source-owning authority
High

Latest recorded AQI is 207

46% confidence · 89 evidence items
Decide
Manesar MCG / relevant source-owning authority
High

Latest recorded AQI is 212

46% confidence · 67 evidence items
Decide
Vikas Sadan MCG / relevant source-owning authority
High

Latest recorded AQI is 207

46% confidence · 67 evidence items
Decide
TERI Gram MCG / relevant source-owning authority
High

Latest recorded AQI is 207

46% confidence · 16 evidence items
Decide

Evidence ledger

Source, time, quality and provenance retained for every signal.

A mapped asset is a place to verify, not a finding and not an emission source. Facility presence is never evidence of pollution. Nothing here has been field-verified — every asset was bulk-imported and the app has no inspection step that could mark one verified, so the right-hand column shows where the record came from instead, which is what should change how far you trust it.

no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
no timestamp
Latest 40 items. Original source records are unchanged.

Operational memory

Reviewed lessons from measured interventions.

No lessons promoted

A lesson is only recorded after an approved action has been carried out and measured. The system does not learn silently from unreviewed outcomes.

Outcome capture · Close the learning loop

Record a measured outcome against an approved proposal, then promote a reviewed lesson.

No outcome to record yet

An outcome is recorded against an approved proposal once its work order has been carried out and measured. Approve something in the queue above first.