Insurance
26 columns
100 rows
8 formats

Claim Fraud Detection Dataset

This synthetic insurance claim fraud detection dataset contains detailed records of claims, including incident specifics, claimant demographics, policy details, and fraud indicators. Designed for developing and testing machine learning models, it enables insurers and researchers to identify patterns of fraudulent activity and improve risk assessment strategies.

opened 9 times
last updated 12 Jul 2025
by GoMask

The brief that made it

Training and evaluating claim fraud detection algorithms

Sample rows

preview · 8 of 100 rows · all 26 columns

claim_idstring	claim_statusstring	fraud_probability_scorefloat	fraud_reportedboolean	claimant_address_countrystring	policy_idstring	claimant_idstring	claim_datedate	incident_datedate	incident_typestring	incident_descriptionstring	claim_amountfloat	claimant_ageinteger	claimant_genderstring	claimant_occupationstring	claimant_address_streetstring	claimant_address_citystring	claimant_address_statestring	claimant_address_postal_codestring	policy_typestring	policy_start_datedate	policy_end_datedate	number_of_previous_claimsinteger	number_of_claims_on_policyinteger	police_report_filedboolean	witness_presentboolean
C10001	under_review	0.13	false	USA	P1023	CLM0001	2023-11-15	2023-11-13	accident	Rear-end collision at intersection.	4200.5	42	female	engineer	12 Cedar Ave	Denver	CO	80203	auto	2021-06-01	2024-06-01	1	2	true	true
C10002	submitted	0.07	false	USA	P1045	CLM0002	2023-10-04	2023-10-03	theft	Bicycle stolen from garage overnight.	650	29	male	teacher	88 5th St	Portland	OR	97209	home	2020-03-12	2025-03-11	0	1	true	false
C10003	approved	0.21	false	USA	P1098	CLM0003	2024-01-12	2023-12-29	fire	Kitchen fire caused by oil spill; minor damage.	3200	36	female	chef	120 Maple Dr	Des Moines	IA	50309	home	2022-07-01	2025-07-01	2	2	false	false
C10004	paid	0.03	false	USA	P1103	CLM0004	2022-07-21	2022-07-20	accident	Fender bender in parking lot.	1200.75	51	male	lawyer	73 Willow Rd	Cleveland	OH	44114	auto	2019-12-01	2024-12-01	0	1	false	false
C10005	approved	0.16	false	USA	P1200	CLM0005	2023-03-05	2023-03-02	natural_disaster	Windstorm damaged roof shingles.	4800.9	60	female	retired	400 Elmwood Blvd	Tampa	FL	33602	home	2020-02-15	2025-02-15	1	1	false	false
C10006	submitted	0.08	false	USA	P1320	CLM0006	2024-04-19	2024-04-17	accident	Car hit a deer on highway.	3200	34	male	accountant	11 Spruce Ct	Madison	WI	53703	auto	2023-03-30	2025-03-30	0	1	true	false
C10007	approved	0.09	false	USA	P1401	CLM0007	2023-12-22	2023-12-20	other	Lost luggage during travel.	850	40	female	consultant	700 Lake Dr	Austin	TX	78701	other	2021-09-01	2024-09-01	0	1	false	false
C10008	paid	0.06	false	USA	P1550	CLM0008	2022-02-14	2022-02-13	accident	Slipped on icy sidewalk, sprained ankle.	2100.25	56	male	postal worker	49 Ridge Way	Buffalo	NY	14202	health	2018-01-01	2025-01-01	3	4	false	true
C10009	under_review	0.12	false	USA	P1074	CLM0009	2023-05-30	2023-05-28	accident	Minor car accident at traffic signal.	1300.5	27	male	sales associate	152 Main St	Memphis	TN	38103	auto	2022-10-01	2025-10-01	1	2	false	true
C10010	approved	0.14	false	USA	P1488	CLM0010	2024-03-20	2024-03-18	fire	Small fire in living room, extinguished quickly.	2800	38	female	graphic designer	90 Oak St	Nashville	TN	37219	home	2021-05-01	2025-05-01	0	1	true	false
C10011	rejected	0.76	true	USA	P1166	CLM0011	2023-09-11	2023-09-10	theft	Wallet stolen at shopping mall.	250	30	female	student	15 Birch Pl	Omaha	NE	68102	other	2022-01-01	2023-12-31	0	1	true	false
C10012	approved	0.05	false	USA	P1289	CLM0012	2023-08-07	2023-08-01	accident	Car skidded on wet road, minor damage.	950	48	male	delivery driver	99 Cherry St	Raleigh	NC	27601	auto	2019-08-01	2024-08-01	2	3	false	true
C10013	approved	0.11	false	USA	P1111	CLM0013	2022-11-23	2022-11-19	fire	Candle caused small fire in bedroom.	1800	52	female	nurse	33 North Dr	Columbus	OH	43215	home	2021-05-15	2025-05-15	0	1	false	false
C10014	paid	0.04	false	USA	P1197	CLM0014	2023-06-14	2023-06-12	accident	Bike accident, minor injuries.	500	22	male	student	321 8th Ave	Seattle	WA	98101	health	2021-08-15	2025-08-15	0	1	false	true
C10015	under_review	0.19	false	USA	P1057	CLM0015	2024-02-08	2024-02-07	natural_disaster	Flooding damaged basement.	7600	47	female	manager	200 Lakeview Dr	St. Louis	MO	63101	home	2020-01-01	2026-01-01	1	1	true	false
C10016	approved	0.13	false	USA	P1158	CLM0016	2023-07-09	2023-07-07	theft	Jewelry stolen from bedroom drawer.	1800	61	female	retired	16 South St	Albany	NY	12207	home	2017-02-01	2024-02-01	2	2	true	false
C10017	approved	0.15	false	USA	P1172	CLM0017	2023-05-18	2023-05-17	accident	Car hit by falling tree branch.	2200	44	male	architect	871 Oakwood Rd	Louisville	KY	40202	auto	2018-11-21	2025-11-21	0	1	false	false
C10018	paid	0.11	false	USA	P1234	CLM0018	2023-08-29	2023-08-26	accident	Fell while hiking, fractured wrist.	1800	32	female	nurse	101 Pine St	Boise	ID	83702	health	2022-03-01	2026-03-01	0	1	false	true
C10019	approved	0.18	false	USA	P1322	CLM0019	2024-01-04	2023-12-31	natural_disaster	Tornado destroyed shed.	2900	46	male	electrician	2 Aspen Ln	Tulsa	OK	74103	home	2021-08-01	2025-08-01	1	1	false	false
C10020	rejected	0.82	true	USA	P1600	CLM0020	2022-12-11	2022-12-09	theft	Laptop stolen from car.	1200	31	male	analyst	900 Adams St	Chicago	IL	60606	auto	2021-09-08	2024-09-08	1	1	true	false

What the 100 rows show

from the 100-row sample

Rejected (claim status) stands out: 8 of its 11 rows have fraud_reported = true, against 2 of 89 for the rest.

10%fraud_reported = true
0.13median fraud_probability_score
4claimant genders
4policy types
5incident types
1,225median claim_amount

Fraud reported rate by claim_statusfraud_reported = true

fraud_probability_score100 rows, in bands of 0.1

Every row with fraud_probability_score of 0.6 or more has fraud_reported = true. None below it.

claimant_address_country100 rows · 7 values

USA56
New Zealand19
Australia10
Canada5
UK5
Ireland4
Austria1

26 columns by typefrom the column list below

string 14
integer 3
float 2
date 4
boolean 3

Columns

26 columns in four groups

blueprint · 26 columns
column	type	description	example
Text 14 columns
`claim_id`	string	Unique identifier for each insurance claimunique	C10001
`policy_id`	string	Unique identifier for the insurance policy associated with the claim	P1023
`claimant_id`	string	Unique identifier for the individual or entity submitting the claim	CLM0001
`incident_type`	string	Type of incident (e.g., theft, accident, fire, natural disaster)theft · accident · fire · natural_disaster · other	accident
`incident_description`	string	Detailed description of the incidentoptional	Tornado destroyed shed.
`claim_status`	string	Current status of the claim (e.g., submitted, under_review, approved, rejected, paid)submitted · under_review · approved · rejected · paid	under_review
`claimant_gender`	string	Gender of the claimantmale · female · other · unknown · optional	female
`claimant_occupation`	string	Occupation of the claimantoptional	engineer
`claimant_address_street`	string	Street address of the claimantoptional	12 Cedar Ave
`claimant_address_city`	string	City of the claimant's addressoptional	Denver
`claimant_address_state`	string	State or province of the claimant's addressoptional	CO
`claimant_address_postal_code`	string	Postal code of the claimant's addressoptional	80203
`claimant_address_country`	string	Country of the claimant's address7 countries · optional	USA
`policy_type`	string	Type of insurance policy (e.g., auto, home, life, health)auto · home · life · health · other	auto
Numbers 5 columns
`claim_amount`	float	Amount being claimed for reimbursement or coverage0 or more	4200.5
`fraud_probability_score`	float	Probability score (0-1) indicating the likelihood of the claim being fraudulent0 to 1 · optional	0.13
`claimant_age`	integer	Age of the claimant at the time of claim0 or more · optional	42
`number_of_previous_claims`	integer	Number of previous claims made by the claimant0 or more · optional	1
`number_of_claims_on_policy`	integer	Total number of claims made against this policy0 or more · optional	2
Dates and times 4 columns
`claim_date`	date	Date when the claim was filed	2023-11-15
`incident_date`	date	Date when the incident leading to the claim occurred	2023-11-13
`policy_start_date`	date	Date when the policy coverage startedoptional	2021-06-01
`policy_end_date`	date	Date when the policy coverage ended or will endoptional	2024-06-01
True or false 3 columns
`fraud_reported`	boolean	Indicates if the claim was flagged or reported as potentially fraudulent	false
`police_report_filed`	boolean	Indicates if a police report was filed for the incidentoptional	true
`witness_present`	boolean	Indicates if a witness was present at the time of the incidentoptional	true

Use it for

An insurance dashboard
The fraud_reported rate, fraud_probability_score by claim_status and a breakdown of claimant_address_country. Excel, Power BI or Tableau.
Why do 10 of 100 rows have fraud_reported = true?
A root-cause class exercise
Hand out the rows and one question. The answer is in the data, not in the brief.
Claims100
C100010.13under_re…
C100110.76rejected
C100200.82rejected
A software demo
Believable claims with policy_id, claimant_id and claim_date to fill a screen in front of a buyer.

This dataset100 rows26 columns

Yours1,000 rows26 columnsclaimant_address_street: UK only

blueprint · claim-fraud-detection-dataset

Behind this dataset

Same schema. As many rows as you need.

These 100 rows came out of a blueprint — 26 columns with generation rules behind each one. Open it in Data Factory to retune a column, add your own, wire in foreign keys, and run it at the size you actually need.

Rules it was built with

Claims missing documentation are denied
Amounts over $10,000 require manual review
Repeat claimants flagged for investigation
Inconsistent accident dates trigger alerts

Rows1,000 rows10,000 rows100,000 rows1,000,000 rows

Open the blueprint in Data Factory

1 credit per row. Preview 20 rows free. Buy rows for the full run.

Exports: CSV, JSON, JSONL, Parquet, SQL, Excel, TSV, XML
Licence: yours to use, including commercially
API slug: claim-fraud-detection-dataset

What should your data show?

No signup. No card.

Sample rows

What the 100 rows show

Columns

Use it for

An insurance dashboard

A root-cause class exercise

A software demo

Make it yours.

Same schema. As many rows as you need.

More in Insurance

What should your data show?