Financial Transaction Fraud Scores

This dataset provides detailed, structured records of bank transactions, each annotated with a synthetic fraud risk score and binary fraud label. It includes transaction, account, merchant, and contextual information, making it ideal for training and evaluating fraud detection and risk assessment models. The dataset supports feature engineering for advanced analytics and machine learning applications in financial security.

  • opened 8 times
  • last updated 29 Jul 2025
  • by GoMask
The brief that made it

Training supervised machine learning models for fraud detection

Sample rows

preview · 8 of 200 rows · all 19 columns
transaction_idstringtransaction_typestringfraud_scorefloatis_internationalbooleanmerchant_category_codestringaccount_idstringtransaction_datetimedatetimetransaction_amountfloatcurrencystringmerchant_idstringmerchant_namestringlocation_citystringlocation_statestringlocation_countrystringchannelstringis_fraudbooleandevice_idstringcustomer_ageintegercustomer_genderstring
TXN1000001purchase0.05false5732ACC0012021-07-11T10:34:17Z52.47USDMRC7192Best ElectronicsSan FranciscoCAUSPOSfalseDEV_SF100129male
TXN1000002purchase0.09false5814ACC0022023-02-19T17:12:49Z18.25USDMRC7811Urban CoffeeBostonMAUSPOSfalseDEV_BOS20322female
TXN1000003purchase0.13false5411ACC0032022-08-03T22:56:21Z97.5EURMRC3620Supermarkt BerlinBerlinBEDEPOSfalseDEV_DEB1434female
TXN1000004transfer0.19trueblankACC0042020-12-31T16:44:00Z25000CHFblankblankZurichZHCHonlinefalseDEV_ZH098256male
TXN1000005refund0.06false5942ACC0052024-01-15T08:25:12Z75GBPMRC4821BookWorld UKLondonLDNGBonlinefalseDEV_LON741female
TXN1000006deposit0.01falseblankACC0062022-03-27T13:02:55Z4000USDblankblankChicagoILUSbranchfalseblank65male
TXN1000007purchase0.1true5411ACC0072023-11-21T19:31:45Z120.4JPYMRC3785Tokyo MarketTokyo13JPPOSfalseDEV_TYO1836male
TXN1000008withdrawal0.08falseblankACC0082022-05-15T07:12:31Z500USDblankblankMiamiFLUSATMfalseblank53female

What the 200 rows show

from the 200-row sample

Transfer (transaction type) stands out: mean fraud_score is 0.28, against 0.08 for the rest.

  • 24%is_international = true
  • 0.08median fraud_score
  • 4customer genders
  • 5channels
  • 9currencies
  • 12location countries
Mean fraud_score by transaction_type200 rows
00.510.10wit…16 r…0.07dep…19 r…0.28tra…24 r…0.12pay…14 r…0.08pur…82 r…0.06ref…8 ro…0.05fee26 r…0.06oth…11 r…
fraud_score200 rows, in bands of 0.1
065130122638410000200.51fraud_score →

Median 0.08, from 0.0 to 0.99.

merchant_category_code130 rows with a value · 70 left blank
  1. 541126
  2. 601217
  3. 573214
  4. 601111
  5. 594210
  6. 58128
  7. 56517
  8. 59414
  9. 49004
  10. 58143
19 columns by typefrom the column list below
  • string 13
  • integer 1
  • float 2
  • datetime 1
  • boolean 2

Columns

19 columns in four groups
blueprint · 19 columns
columntypedescriptionexample
Text 13 columns
transaction_idstringUnique identifier for each bank transaction record.uniqueTXN1000001
account_idstringUnique identifier for the account involved in the transaction.ACC001
currencystringISO 4217 currency code for the transaction (e.g., USD, EUR, GBP).9 currenciesUSD
transaction_typestringType of transaction (e.g., withdrawal, deposit, transfer, payment, purchase).8 valuespurchase
merchant_idstringUnique identifier for the merchant or counterparty, if applicable.optionalMRC7192
merchant_namestringName of the merchant or counterparty, if available.optionalBest Electronics
merchant_category_codestringMerchant category code (MCC) indicating the type of merchant.optional5732
location_citystringCity where the transaction took place, if available.optionalSan Francisco
location_statestringState or province where the transaction took place, if available.optionalCA
location_countrystringCountry where the transaction took place, if available.12 countries · optionalUS
channelstringChannel through which the transaction was initiated (e.g., online, ATM, POS, mobile, branch).6 valuesPOS
device_idstringIdentifier for the device used to initiate the transaction, if available.optionalDEV_SF1001
customer_genderstringGender of the account holder, if available.male · female · other · unknown · optionalmale
Numbers 3 columns
transaction_amountfloatMonetary amount of the transaction in the account's currency.0 or more52.47
fraud_scorefloatSynthetic fraud risk score for the transaction, typically ranging from 0.0 (no risk) to 1.0 (high risk).0 to 10.05
customer_ageintegerAge of the account holder at the time of transaction, if available.0 or more · optional29
Dates and times 1 column
transaction_datetimedatetimeDate and time when the transaction occurred.2021-07-11T10:34:17Z
True or false 2 columns
is_internationalbooleanIndicates if the transaction is international (true) or domestic (false).false
is_fraudbooleanLabel indicating whether the transaction is fraudulent (true) or not (false).false

Use it for

  • is internation…24%48 of 200 rowsmean fraud score by t…0.10with…0.07depo…0.28tran…0.12paym…

    A finance dashboard

    The is_international rate, fraud_score by transaction_type and a breakdown of merchant_category_code. Excel, Power BI or Tableau.

  • Why do the 24 transfer rows have a mean fraud_score of 0.28?

    A root-cause class exercise

    Hand out the rows and one question. The answer is in the data, not in the brief.

  • A software demo

    Believable transactions with account_id, transaction_datetime and transaction_amount to fill a screen in front of a buyer.

Not quite right?

Make it yours.

Same 19 columns, your size and your rules. See 20 rows before you pay.

Preview 20 rows free

1,000 rows: $1.99 · 10,000: $12.99One-time. No subscription. All prices

This dataset200 rows19 columns
Yours1,000 rows19 columnslocation_city: UK only

blueprint · financial-transaction-fraud-scores

Behind this dataset

Same schema. As many rows as you need.

These 200 rows came out of a blueprint — 19 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
  • Each transaction has unique account and merchant IDs
  • Fraudulent transactions represent 2% of dataset
  • High-value transactions occur less frequently
  • Transaction geography provided at city level
Rows
Open the blueprint in Data Factory

1 credit per row. New accounts start with 25 free credits.

Exports
CSV, JSON, JSONL, Parquet, SQL, Excel, TSV, XML
Licence
yours to use, including commercially
API slug
financial-transaction-fraud-scores

More in Finance

Browse all
Also in FinanceFinancial Transaction Fraud PatternsFinance Expense Categorization DatasetPersonal Finance Budgeting RecordsAll 1,954 datasets

What should your data show?

Preview 20 rows free
No signup. No card.