Validation

No black box. Every number has a source.

Vigilo's scoring is deterministic: the same inputs always produce the same outputs. Every signal links back to its original official publication. We pre-register, backtest and publish — including the results we're not proud of.

How scoring works

Deterministic. Traceable. Reproducible.

Seven domain scores, one composite. Each sub-score is computed from a fixed formula over the last 30 days of verified data. No machine learning, no hidden weights.

Step 01

Ingest

44 verified feeds polled every 15 minutes. Raw events are deduplicated, geolocated and timestamped.

Step 02

Score

Each event contributes a severity point to its domain (0–100). ERA5 25-year baseline adjusts for seasonal norms.

Step 03

Composite

Seven domain scores are combined with fixed weights into a single 0–100 composite. Weights are published.

Step 04

Cite

Every score response includes the source list that produced it. Click any source to see the original publication.

Composite modifier

Structural fragility

The composite is the live, multi-domain signal — but it is modified by structural fragility drawn from the INFORM Risk Index (European Commission JRC and UN OCHA), the humanitarian sector's standard country-risk measure. Fragility lifts a quiet but fragile country to at most the 'low' band, and amplifies the live signal of the most fragile countries by up to 20%. It never dominates: a stable country stays low and the top of the table is always driven by live events. Each country's record carries both the live score and the fragility applied, so the structural contribution is fully traceable.

Backtests

We tested. We published. Including the failures.

Before adding a new signal source, we run retrospective validation against historical events. Results — good and bad — are published on GitHub.

PASSED

Dengue pre-season

Thailand 2023. Scored 71/100 two weeks before the official WHO declaration. Seasonal climate signal correctly elevated prior to outbreak confirmation.

Lead time +14 days · confidence 0.71 · Backtest 1/2
Full report ↗
PASSED

Conflict escalation

Sudan 2023. GDELT composite flagged 14 days before major media coverage. Conflict domain score crossed alert threshold ahead of international reporting.

Lead time +14 days · PR-AUC +0.081 · Backtest 1/2
Full report ↗
FAILED

Conflict compound

Sudan 2023 phase 2. Combined bio + conflict signal. With more data and six countries, the combined model decisively loses to the seasonal baseline. The predictive hypothesis is not supported.

PR-AUC 0.0724 vs 0.1925 · skill −0.1201 · Backtest 2/2
Full report ↗
Signal sources

44 verified feeds

Every feed is manually reviewed before inclusion. Feeds are re-evaluated quarterly. If a source degrades, it is removed — not quietly downweighted.

WHO IHRInternational Health Regulations alerts
WHO DONDisease Outbreak News
WHO AFROAfrica regional office
WHO EMROEastern Mediterranean region
WHO EUROEuropean region surveillance
CDC MMWRMorbidity & Mortality Weekly
ECDC CDTREU communicable disease report
Africa CDCAfrican Union health surveillance
ProMEDProgram for Monitoring Emerging Diseases
FluNetGlobal influenza surveillance
Nigeria NCDCNigeria Centre for Disease Control
GDACSGlobal Disaster Alert & Coordination
USGSSeismic & volcanic hazards
NASA EONETEarth observatory event tracker
NOAAWeather, storm & ocean events
ReliefWebOCHA humanitarian situation reports
WFP HungerMapWorld Food Programme live hunger data
OpenDengueArboviral disease time series
ACLEDArmed Conflict Location & Event Data
ERA525-year ECMWF climate reanalysis
OpenAQOpen air quality measurements
CopernicusEU earth observation service
FDA recallsUS food & drug safety alerts
EFSAEuropean Food Safety Authority
IATATravel advisories & restrictions
Limits

What we don't do

Knowing what a system doesn't do is as important as knowing what it does.

No ML inference

Scores are computed from a fixed deterministic formula. No model weights, no training data, no drift. The same inputs always produce the same output — forever.

No invented numbers

Every figure comes from a named official publication. If we can't source it, we don't score it. No estimates, no imputed values, no synthetic signals.

No proprietary lock-in

The methodology document is public. The backtest code is on GitHub. You can reproduce any score independently. We do not claim accuracy we can't demonstrate.