claritystorm/fema-nfip-flood-insurance-claims

---
license: other
license_name: public-domain
task_categories:
- tabular-classification
- tabular-regression
tags:
- insurance
- flood
- climate-risk
- fema
- nfip
- real-estate
- geospatial
- united-states
pretty_name: FEMA NFIP Flood Insurance Claims 1978-Present
size_categories:
- 1M<n<10M
---

# FEMA NFIP Flood Insurance Claims 1978–Present

2 million+ paid National Flood Insurance Program (NFIP) claims since 1978 — cleaned, enriched, and geocoded.
Includes flood zone classifications, coverage amounts, building characteristics, damage breakdowns, and **ClarityStorm-computed ZIP-level risk scores**.
The essential dataset for flood risk modeling, climate finance, and real estate AI.

| 📊 Records | 📅 Coverage | 🏷️ License | 🔄 Updated |
|-----------|-------------|-----------|-----------|
| 2M+ paid claims | 1978–present (46 years) | Public Domain | Annual |

**This repo contains a free 1,000-row sample.**
Full dataset (CSV + Parquet) → **[claritystorm.com/datasets/fema-flood-insurance](https://claritystorm.com/datasets/fema-flood-insurance)**

---

## Quick Start

```python
from datasets import load_dataset
import pandas as pd

# Load the 1,000-row sample
ds = load_dataset("claritystorm/fema-nfip-flood-insurance-claims")
df = ds["train"].to_pandas()

# Total claims paid by state
print(df.groupby("property_state")["total_amount_paid"]
.sum().sort_values(ascending=False).head(10))

# Average claim payout by flood zone risk category
print(df.groupby("flood_zone_category")["total_amount_paid"]
.mean().sort_values(ascending=False).round(0))

# ZIP-level risk scores (ClarityStorm-computed)
high_risk = df[df["zip_risk_score"] >= 80]
print(f"High-risk ZIPs (score ≥80): {high_risk['reported_zip_code'].nunique()}")

# Decade-over-decade damage trend
print(df.groupby("decade_of_loss")["total_amount_paid"].mean().round(0))
```

## Use Cases

- **Flood risk modeling** — 46-year claims history with flood zone, building type, and damage breakdowns
- **Climate finance & physical risk** — ZIP-level risk scores, damage trends, and coverage ratio features
- **Real estate AI** — flood risk as a feature for property valuation and mortgage underwriting
- **Insurance loss estimation** — historical loss distributions by flood zone, state, and occupancy type
- **Catastrophe (CAT) model validation** — compare model outputs against actual NFIP paid claim distributions
- **FEMA policy analysis** — post-FIRM construction trends, elevated building performance, and NFIP reform research

## Schema (selected fields)

| Field | Type | Description |
|-------|------|-------------|
| claim_id | string | Unique claim identifier |
| year_of_loss | int | Year the flood loss occurred |
| property_state | string | State abbreviation (2-letter) |
| reported_zip_code | string | 5-digit ZIP code of insured property |
| latitude | float | Property latitude (where available) |
| longitude | float | Property longitude (where available) |
| flood_zone | string | FEMA flood zone designation (e.g. AE, VE, X) |
| flood_zone_category | string | Simplified risk category (High Risk SFHA, Moderate Risk, etc.) |
| occupancy_type_label | string | Human-readable occupancy type |
| total_building_coverage | float | Building insurance coverage limit ($) |
| total_amount_paid | float | Total claim payment (building + contents + ICC) |
| coverage_ratio | float | Claim payout as fraction of coverage limit |
| decade_of_loss | int | Decade of loss year (e.g. 1990, 2000, 2010) |
| zip_claim_count | int | Total NFIP claims from this ZIP code |
| zip_avg_claim | float | Average total claim payout for this ZIP |
| zip_risk_score | float | ZIP-level risk score 1–100 (frequency × severity) |

## ⬇️ Get the Full Dataset

| Tier | Price | Includes |
|------|-------|----------|
| Sample | Free | 1,000 rows, Public Domain (this repo) |
| Complete | $99 | Full 2M+ rows, CSV + Parquet, commercial license |
| Annual | $199/yr | Complete + annual updates |

👉 **[Purchase at claritystorm.com/datasets/fema-flood-insurance](https://claritystorm.com/datasets/fema-flood-insurance)**

## Source

Federal Emergency Management Agency (FEMA), OpenFEMA — FIMA NFIP Redacted Claims v2.
FEMA OpenFEMA data is a US federal government work in the **public domain** (17 U.S.C. 105).
All personally identifiable information has been redacted by FEMA.

Processed and enriched by [ClarityStorm Data](https://claritystorm.com).