How GeoGridIQ turns evidence into reviewable outage-risk intelligence

What GeoGridIQ estimates

GeoGridIQ estimates outage risk for a declared geography and future time window. It combines only source evidence authorized for that region, horizon, and output track, freezes what was genuinely available at issue time, and evaluates the snapshot through an exact verified public-serving path. If that path cannot be verified, the public result is unavailable. It is designed to support monitoring and prioritization, not to act as an official outage notice, emergency warning, or guarantee about a particular asset or customer.

What the result means

A GeoGridIQ output is more than a score. Its contract includes the target outcome, region or grid cell, issue time, valid-from and valid-to times, forecast horizon, prediction mode, source freshness, confidence or evidence-quality context, explanatory drivers, and current availability state. These fields protect against common misreadings. Probability is not confidence. Relative risk is not automatically a calibrated probability. Severity describes conditions, consequence describes possible impact, and a top driver is an associated input—not proof of cause.

Grid-cell and regional results describe only their named analysis unit. They do not identify an exact line, pole, transformer, address, or customer. Critical-infrastructure context may raise the importance of human review, but proximity alone does not make an outage physically more likely.

How evidence becomes a forecast

The workflow starts by defining the forecast question. GeoGridIQ then gathers permitted outage, weather, vegetation, geographic, exposure, and historical evidence; records provider, product, timestamps, coverage, and quality; aligns different geometries and time windows; and freezes a versioned feature snapshot at issue time. Runtime controls resolve the exact region-and-horizon model, verify artifact and feature compatibility, generate and persist the intended prediction batch, and apply separate trust, freshness, validity, and display gates.

This process is deliberately fail-closed. A model file without a current, unexpired batch is not a current forecast. A missing outage feed is unknown, not zero active outages. Missing vegetation evidence is not zero vegetation risk. GeoGridIQ never substitutes another region or horizon's model or presents stale output as current. Fallback generation or scoring may exist under its own contract, but public capability remains unavailable until a separate public-serving fallback contract is implemented and verified.

Why time and validation matter

A prediction may use only evidence that was available by its declared issue-time cutoff. Later outage reports, final weather observations, restorations, and source revisions belong to validation, not the original forecast. GeoGridIQ stores the forecast before the outcome and later evaluates it under declared spatial, temporal, and provider-coverage rules. When observation coverage is incomplete, an unmatched prediction may be unknown rather than a confirmed false positive.

How to use it responsibly

A reviewer should confirm the valid window, source state, verified public prediction mode, confidence and missingness, associated drivers, and consequence context. The result should be compared with official weather alerts and utility outage information, then interpreted through applicable safety procedures and field expertise. GeoGridIQ does not replace ECCC warnings, utility SCADA or OMS, field inspection, emergency procedures, or authorized operational judgment.

Availability changes over time. Consult the current capability table for the exact region, horizon, source, model, and batch state as of its displayed UTC timestamp.