🌊 OpenFEMA NFIP Flood Insurance Claims

Pricing

Rating

Developer

Actor stats

Categories

Share

Pull redacted, claim-level National Flood Insurance Program (NFIP) loss records straight from FEMA's OpenFEMA platform — the authoritative, government-published history of what the United States actually pays out when it floods. Filter by state and year-of-loss and get one structured JSON record per claim: flood zone, building and contents amounts paid, the flood event, and the loss date. This is the granular paid-loss data that P&C insurers, reinsurers, catastrophe modelers, and climate-risk analysts use to calibrate flood pricing and exposure — sourced from the only dataset that covers the entire federal flood program rather than a single carrier's book.

One actor. One token. The claim-level flood-loss history behind cat-model calibration and flood-risk pricing — delivered as pay-per-record JSON, no enterprise data licence.

Why use this

• Actual paid losses, not modeled estimates. Most flood-risk products sell you a model's prediction. This is FEMA's record of what was actually paid on real NFIP claims — the ground-truth your model gets calibrated against.
• Whole-program coverage. The NFIP underwrites the large majority of residential flood policies in the US, so this is a near-complete federal flood-loss picture, not one carrier's slice.
• Building vs contents split. Each claim separates amountPaidOnBuildingClaim and amountPaidOnContentsClaim — essential for severity modeling and coverage-mix analysis.
• Flood-zone tagged. Every claim carries its ratedFloodZone, letting you analyze loss severity by zone (A/AE/VE/X) — the backbone of flood-rating work.
• Official source, ETL-ready. Reads FEMA's public OpenFEMA dataset; redacted for privacy at the source. Stable flat JSON loads straight into Snowflake, BigQuery, Postgres, or a notebook.

What you get

Each record represents one redacted NFIP flood-insurance claim. Core fields:

• state — two-letter state of the insured property
• dateOfLoss — the date the flood loss occurred
• ratedFloodZone — the FEMA flood zone the policy was rated in (e.g. AE, VE, X, A)
• amountPaidOnBuildingClaim — USD paid on the building portion of the claim
• amountPaidOnContentsClaim — USD paid on the contents portion of the claim
• floodEvent — the named flood event associated with the loss (when designated)

Records are redacted at the FEMA source to protect policyholder privacy — there is no name, full address, or policy number, by design.

Use cases

• Flood-risk pricing & rate filing — P&C actuaries aggregate amountPaidOnBuildingClaim and amountPaidOnContentsClaim by ratedFloodZone and state to derive empirical loss-cost relativities for private-flood rate filings competing with the NFIP.
• Catastrophe-model calibration — Cat modelers and reinsurers backtest vendor flood models (against the actual NFIP paid-loss distribution by zone and event) to validate severity curves before binding treaty capacity.
• Reinsurance & ILS underwriting — Reinsurance and insurance-linked-securities desks size flood exposure and event severity (e.g. by pulling all claims tagged to a named hurricane's floodEvent) to price retrocession and cat-bond layers.
• Climate-risk analytics — Climate and ESG analysts trend flood-loss frequency and severity by state and year-of-loss to quantify physical-risk exposure for real-estate, mortgage, and municipal-bond portfolios.
• Mortgage & real-estate risk — Lenders and proptech firms join paid-loss intensity by zone to property portfolios to flag concentration risk and inform flood-insurance requirements.
• Public-policy & academic research — Researchers study NFIP loss patterns, repetitive-loss properties, and the equity of flood-zone rating using the program's own paid-claim record.
• Disaster-response planning — Emergency-management and resilience teams quantify historical loss footprints of major flood events to prioritize mitigation spend.

Sample output

{
"state":"TX",
"dateOfLoss":"2017-08-27",
"ratedFloodZone":"AE",
"amountPaidOnBuildingClaim":184320.55,
"amountPaidOnContentsClaim":41875.00,
"floodEvent":"Hurricane Harvey"
}

Reading the data: flood zones, building vs contents, and named events

Three fields do most of the analytical work, and it helps to understand how they interact:

• ratedFloodZone is the FEMA flood zone the policy was rated in. The high-risk Special Flood Hazard Areas (A, AE, and the coastal high-velocity VE) carry the highest expected severity; X zones are moderate-to-low risk and tend to show smaller, less frequent paid losses. Grouping amountPaidOnBuildingClaim by ratedFloodZone is the single most common analysis this dataset supports — it produces empirical severity relativities by zone.
• amountPaidOnBuildingClaim vs amountPaidOnContentsClaim must be analyzed separately, not summed blindly. NFIP underwrites and adjusts building and contents coverage independently, and the mix varies sharply by occupancy (a slab-on-grade home vs a contents-heavy small business). Severity curves built on the building component behave very differently from contents.
• floodEvent ties a claim to a named event when one is designated. This is what lets you reconstruct the loss footprint of a specific catastrophe — pull a state and year window, then group on floodEvent to isolate, say, every claim attributed to a single landfalling hurricane.

Common analysis recipes

• Single-event severity: {"state": "TX", "yearOfLossFrom": 2017, "maxResults": 5000} then group returned rows on floodEvent to isolate one hurricane's claims and chart the paid-loss distribution by ratedFloodZone.
• Zone relativity table: pull a state's recent years and pivot mean/percentile amountPaidOnBuildingClaim by ratedFloodZone — the empirical input to a private-flood rate filing.
• Coastal vs inland mix: compare VE/AE (coastal) against A/X (inland) severity to size storm-surge vs riverine exposure in a book.
• Trend over time: run the actor for successive yearOfLossFrom cuts to trend frequency and severity and quantify a warming-climate signal in the paid-loss record.

Input parameters

How to use

Python (apify-client)

from apify_client import ApifyClient
client = ApifyClient("YOUR_TOKEN")
run = client.actor("nexgendata/openfema-nfip-flood-claims").call(run_input={
"state":"TX",
"yearOfLossFrom":2017,
"maxResults":1000,
})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
print(item["dateOfLoss"], item["ratedFloodZone"],
 item["amountPaidOnBuildingClaim"], item["floodEvent"])

cURL

curl-X POST "https://api.apify.com/v2/acts/nexgendata~openfema-nfip-flood-claims/run-sync-get-dataset-items?token=YOUR_TOKEN"\
-H"Content-Type: application/json"\
-d'{
 "state": "FL",
 "yearOfLossFrom": 2020,
 "maxResults": 5000
 }'

No-code (Zapier / Make)

Use the Apify "Run actor and get dataset items" action with actor ID nexgendata/openfema-nfip-flood-claims, then load the rows into a spreadsheet or BI tool and pivot paid losses by ratedFloodZone and state.

Pricing

This actor runs on Apify's pay-per-event (PPE) model — $0.10 per result (one charge per claim record returned), plus a sub-cent actor-start event. No subscription, no seat licence, no minimum.

Worked example:

• A targeted single-event pull returning 250 claims → $25.00
• A state-year analysis returning 500 claims → $50.00
• A multi-year exposure study capped at maxResults=5000 → up to $500.00 (use maxResults to bound cost precisely)

Because billing is per result, a narrow filter that returns few claims costs only for those rows — you never pay for an empty run.

How this compares to a commercial flood-data vendor

Commercial flood-risk and catastrophe-data products (cat-model vendor data feeds, proprietary flood-score APIs, and reinsurance-broker analytics) sell flood-loss intelligence on annual enterprise licences running into the high five and six figures. Those products add modeled scores, parcel-level geocoding, and proprietary hazard layers you may still need. But the actual NFIP paid-loss record — the calibration target underneath all of them — is public FEMA data. If your workflow is calibrating a model, sizing event severity, or trending flood losses by zone and state, you do not need a six-figure data licence to get the claim-level paid amounts. You need this actor, queryable from your own code at pay-per-record cost.

FAQ

Q: Where does the data come from? A: FEMA's OpenFEMA platform — the official, public NFIP redacted claims dataset published by the federal government.

Q: Is personal information exposed? A: No. FEMA redacts the dataset at the source to protect policyholder privacy: there are no names, full addresses, or policy numbers. The records are loss-level facts (zone, amounts, date, event).

Q: What does ratedFloodZone mean? A: It's the FEMA flood zone the policy was rated in — e.g. AE and A (high-risk Special Flood Hazard Areas), VE (coastal high-velocity), and X (moderate-to-low risk). Severity analysis by zone is one of the most common uses of this data.

Q: Why are building and contents split into two fields? A: NFIP policies cover building and contents separately, and they are adjusted and paid separately. Keeping amountPaidOnBuildingClaim and amountPaidOnContentsClaim distinct is essential for accurate severity and coverage-mix modeling.

Q: Can I pull all claims for a specific hurricane? A: Filter by state and yearOfLossFrom to narrow to the relevant period, then filter the returned rows on floodEvent for the named event (e.g. Hurricane Harvey) in your downstream pipeline.

Q: How current is it? A: The dataset reflects FEMA's published NFIP claims as of its latest OpenFEMA refresh. Schedule the actor periodically via Apify's scheduler to pick up newly published claims after major events.

Schema stability & versioning

This actor follows NexGenData's additive-only schema contract. New fields may be added over time (appearing as new JSON keys, defaulting to null for older runs), but existing fields are never renamed or removed without a major-version bump and advance notice. Field semantics — currency, date formats, zone codes — are never silently changed. Build production pipelines on this output with confidence.

Compliance & legal

• This actor reformats public, redacted FEMA OpenFEMA data. It is not affiliated with, endorsed by, or sponsored by FEMA or the NFIP.
• The data is redacted at the FEMA source for privacy. Nothing here is insurance, legal, or investment advice.
• For any regulated decision — rate filings, treaty pricing, underwriting — verify against the primary OpenFEMA dataset and apply appropriate actuarial review.
• You are responsible for ensuring your downstream use complies with applicable laws.

Related NexGenData actors

Part of NexGenData's Environmental / ESG and Finance clusters — pair this actor with:

• EPA TRI Toxic Release Facilities
• EPA ECHO Enforcement Scraper
• FDIC Bank Financials & Health
• US Treasury Auction Results
• Global Central-Bank Policy Rates (BIS)

Browse the full 200+ actor catalog at https://apify.com/nexgendata?fpr=2ayu9b.