Synthetic Insurance Claims Dataset
Synthetic insurance claim filings with labelled fraud signals — for testing claim-triage models, anomaly detection and underwriting AI.
Intended use
- Testing
- Validation
- Demonstration
Not intended for
- Real claim decisions
- Profiling real policyholders
Documentation level
High
Overview
Synthetic motor and property claim filings with structured fields covering policyholder profile, incident, loss estimate and labelled fraud indicators. Designed for benchmarking claim-triage models and anomaly detection pipelines without exposing real customer data.
What it contains
- Dataset file (Parquet)
- Sample preview (250 records)
- Data dictionary
- waiser data Dossier
- License terms
- Version log
Generation method
Generated via a constrained statistical pipeline calibrated against published industry distributions. Fraud-flagged records are injected at controlled rates with documented indicator patterns. No real claim files were used as seed.
Permitted and prohibited uses
Permitted
- Benchmarking claim-triage models
- Anomaly detection R&D
- Demonstration of underwriting AI pipelines
- Internal validation exercises
Prohibited
- Real claim adjudication
- Combining with external data to infer real identities
- Use outside the permitted license scope
Dataset structure
| Column | Description |
|---|---|
| claim_id | Unique synthetic claim identifier |
| policy_type | motor / property |
| incident_date | Synthetic incident date |
| loss_estimate | Estimated loss amount |
| fraud_label | 0 / 1 — injected fraud indicator |
| indicator_pattern | Documented synthetic indicator class |
Quality notes
Distributions calibrated to public industry references. Fraud-label base rate is documented and may not match any specific insurer's portfolio.
waiser data Readiness Profile
Ethical notes
Synthetic profiles do not represent real demographic distributions. Do not use as a basis for population-level pricing or risk claims.
License
waiser data Standard License — Defined Use
Full license terms are provided with the dataset delivery. Access is granted after review of intended use.
Version log
ready to request access?
Tell us your intended use and we'll respond with the executed licence, the full dossier and a secure delivery link bound to your request.