Forecast outage risk earlier
Estimate likely outage zones before impacts escalate into broader operational disruption.
AI-assisted utility operations intelligence
Predict grid risk before outages escalate.
GeoGridIQ combines live outage data, weather conditions, vegetation density, infrastructure exposure, an optional XGBoost point forecast agent, and a deterministic rules-based safety layer to help utility teams make faster operational decisions.
The operational gap
Outage response should not begin only after customers lose power.
What is happening?
Why is it happening?
What is likely to happen next?
How confident are we?
What should operations do now?
What is GeoGridIQ?
GeoGridIQ is a utility intelligence platform for outage prediction and operational preparedness.
GeoGridIQ helps utilities and infrastructure teams understand where outage risk is increasing, why it is increasing, how confident the forecast is, and which operational actions should be reviewed. It uses outage data, weather intelligence, vegetation signals, critical infrastructure context, GIS analytics, model validation, and crew readiness signals.
Combine utility and environmental signals
Use weather, NDVI, historical outages, critical assets, and GIS context as decision-support evidence.
Support utility operations teams
Help operators, planners, emergency coordinators, and field supervisors prepare for grid stress.
Platform capabilities
From raw GIS feeds to operational action.
The platform brings live outage context, environmental intelligence, infrastructure exposure, and crew readiness into one decision-support surface.
Live Outage Map
Shows outages, customers affected, causes, status, and severity in a live MapLibre dashboard.
Weather Intelligence
Tracks temperature, precipitation, rain, snowfall, wind speed, and wind gusts from forecast data.
Vegetation Intelligence
Uses Sentinel-2 imagery and NDVI analysis to identify vegetation pressure near infrastructure.
Hybrid Prediction Engine
Uses XGBoost point forecasts when trusted inputs are available, with bounded rules-based scoring as the operational fallback.
Critical Infrastructure Threat Detection
Scores risk near hospitals, substations, water treatment, telecom, emergency, school, highway, and rail assets.
Crew Deployment Optimization
Recommends staging areas, deployment radius, nearest crew, and rough travel time estimates.
Intelligence workflow
Designed to turn operational data into action.
- 01Data Sources
Hydro outages, Open-Meteo, Sentinel-2 NDVI, asset data, and crew context.
- 02Risk Scoring
Weather, vegetation, outage density, and infrastructure exposure are scored together.
- 03Probability + Confidence
Trusted point forecasts use XGBoost; broad map layers and fallback decisions remain explainable and rules-based.
- 04Operational Action
The interface surfaces crew staging, threat protection, and briefing recommendations.
Explainable prediction
Every prediction should answer why.
Operators should not have to trust a black box. The platform explains each risk score using the signals that contributed to the prediction.
RiskSevere
Probability82%
ConfidenceHigh
- Dense vegetationNDVI elevated
- Wind gusts above thresholdWeather stress
- Rainfall detectedSurface loading
- Historical outages nearbyRecurring corridor
- Critical infrastructure exposurePriority asset
Wind activity remains elevated across the Laurentides region.
Vegetation density and historical outage patterns indicate elevated infrastructure stress. Critical assets within the risk zone include one hospital and two telecom sites.
Recommended action
Stage crews near Laval and monitor wind gust escalation over the next 6 hours.
Prediction roadmap
Credible prediction starts with measurement.
The platform uses a hybrid prediction architecture: bounded rules-based intelligence is always available, while a trained XGBoost point forecast agent can be used when an artifact exists and its required signals and confidence safeguards pass.
Rules-based prediction engine
Always available for broad operational surfaces and as the deterministic fallback layer.
XGBoost outage prediction agent
Checking the live model artifact and training metrics.
Calibration and temporal validation
Continue measuring predicted versus observed outages before treating ML output as production-grade intelligence.
Prediction accuracy dashboard
A prediction system should measure itself.
The goal is continuous improvement, not blind automation.
--training samples
--positive samples
--holdout accuracy
--precision
--recall
Hybridsafe fallback mode
Open source technical stack
Built from practical geospatial and operations components.
Backend
Python, Django, PostgreSQL/PostGIS, GeoServer, pg_tileserv, pg_featureserv
Frontend
MapLibre, JavaScript, Chart.js
Infrastructure
Docker, Docker Compose, Terraform, AWS
Data
Hydro-Québec outage data, Open-Meteo, Microsoft Planetary Computer, Sentinel-2 L2A, NDVI analysis
Build before the next storm
Build smarter utility operations before the next storm.
GeoGridIQ helps teams see risk earlier, understand contributing factors, protect critical infrastructure and life-safety assets, and deploy crews before outages escalate.