FEMA NFHL · XGBoost · SHAP · Wildfire

Enrich 10,000 Properties in 5 Minutes

Flood zone, wildfire score, vegetation proximity, slope, and composite risk — with XGBoost + SHAP explainability so underwriters know exactly why each property received its score.

The Problem

Where the time goes today

Underwriters need geospatial answers for every property — flood zone, wildfire proximity, slope, composite risk. Manual lookup is hours per property; thousands of properties per quarter.

Hours per property for manual FEMA NFHL + terrain + vegetation lookup

Subjective risk scoring with no audit trail

Black-box models regulators won't accept

Quarterly portfolio refresh rarely actually quarterly

The Workflows

Insurance & Catastrophe Risk workflows on the platform

Fourteen workflows that take a property table in and put a per-property explainable risk score out.

Workflow Name & what it automates Time saved
#242Property Risk Enrichment
Batch geocode + flood zone + wildfire + slope + composite risk
#243XGBoost + SHAP Risk Scoring
Interpretable risk model with per-feature contribution analysis
#235FEMA NFHL Spatial Join
Automated flood zone lookup (A, AE, X, V, VE) with effective-date stamp
#237Wildfire Risk Scoring
0–100 composite score from vegetation + slope + fuel + WUI proximity
#239Hail Swath Loss Modeling
Historic hail swath intersection with property polygons + loss-curve application
#241Tornado-Track Loss Modeling
Storm Prediction Center track database joined to insured property layer
#244Roof Material from Imagery
AI roof material + condition classification from aerial imagery
#62Flood Risk — Infrastructure
Per-asset SFHA classification with audit-ready trail
The Math

ROI vs. manual processing

Modeled on a mid-sized P&C insurer enriching 250,000 properties per year.

Metric Manual / current tooling GeoDataConverter
Annual enrichment cost$304,375$37,488
Net savings$266,887 (712% ROI)
Processing time (250K properties)3,875 hours12 hours
Score explainabilityNonePer-feature SHAP
Differentiator

SHAP — every score comes with a "why"

The XGBoost model produces SHAP values per property showing exactly which factors contributed to the risk score: flood zone, wildfire proximity, slope, vegetation, roof material, distance to fire station. This is auditable, explainable risk scoring that underwriters and state regulators will accept.

Free 1,000-Property Risk Enrichment

Send a CSV of 1,000 property addresses. Get back flood zone, wildfire score, slope, and composite risk with SHAP explanations.

  • Flood zone (FEMA NFHL effective)
  • Wildfire score (0–100)
  • Slope analysis
  • Composite risk + SHAP per property