Methodology
Scores are explainable, source-aware, and confidence-limited.
OSM/PostGIS base coverage
This stage contributes to either the score, the confidence value, or the evidence trail shown to users and admins.
Public-source extraction
This stage contributes to either the score, the confidence value, or the evidence trail shown to users and admins.
Language normalization
This stage contributes to either the score, the confidence value, or the evidence trail shown to users and admins.
Evidence review
This stage contributes to either the score, the confidence value, or the evidence trail shown to users and admins.
Scoring weights
This stage contributes to either the score, the confidence value, or the evidence trail shown to users and admins.
Backend heatmap precomputation
This stage contributes to either the score, the confidence value, or the evidence trail shown to users and admins.
Confidence and limitations
This stage contributes to either the score, the confidence value, or the evidence trail shown to users and admins.
No fake precision
Sparse regions are intentionally marked as low-confidence area estimates.
Sensitive claims are neutral
The platform reports public risk signals, not personal accusations.
Source-driven confidence
Confidence rises only when map, evidence, and review signals support the claim.