Check the setup
- Python 3.10 or newer
- A notebook environment such as Jupyter or Google Colab
- A two-letter U.S. state code
Redacted National Flood Insurance Program claims for building local flood-loss histories and insurance-signal prototypes.
From source to product signal
OpenFEMA publishes redacted NFIP claim records with occupancy, dates, and paid amounts. Start with one state and a recent year. Claims are insured losses, not a complete flood census, and redaction plus policy changes limit neighborhood-level conclusions.
Install the packages, then run the notebook cell.
python -m pip install pandas requests
import pandas as pd
import requests
response = requests.get(
"https://www.fema.gov/api/open/v2/FimaNfipClaims",
params={
"$filter": "state eq 'FL' and yearOfLoss eq 2024",
"$select": "id,state,countyCode,yearOfLoss,dateOfLoss,occupancyType,amountPaidOnBuildingClaim,amountPaidOnContentsClaim",
"$top": 100,
},
timeout=60,
)
response.raise_for_status()
claims = pd.DataFrame(response.json()["FimaNfipClaims"])
claims["retrieved_at_utc"] = pd.Timestamp.now(tz="UTC")
print(claims.head())Test a useful signal
Test whether a bounded claims extract can power a county flood-loss sketch.
Federal Emergency Management Agency is a government source. Last verified 2026-08-18. Temporal coverage: historical NFIP claims through the current OpenFEMA extract.