Fraud Detection

CSV File Description


The Data folder contains the following data:

  • Untagged_Transactions.csv: transaction data from on an online store
  • Account_Info.csv: Anonomyzed account information
  • Fraud_Transactions.csv: transactions identified as fraud

Untagged_Transactions.csv contain the following fields:

ColumnsTypeDescription
transactionIDStringUnique transaction Id
accountIDStringUnique account Id
transactionAmountUSDDoubleTransaction amount in USD e.g., 12345.00
transactionAmountDoubleTransaction amount in currency expressed in transactionCurrencyCode e.g., 12345.00
transactionCurrencyCodeStringCurrency code of the transaction. 3 alphabet letters, e.g., USD
transactionCurrencyConversionRateDoubleConversion rate to US Dollars, e.g. 1.0000 for USD to USD
responseCodeStringresponse code from card issuer payment authorization
digitalItemCountintegerNumber of digital items purchased. (e.g. music, ebook, software, etc, that can be directly downloaded online)
physicalItemCountintegerNumber of physical items purchased (that needs to be shipped)
purchaseProductTypeStringType of product purchased
shippingAddressStringshipping street address
shippingPostalCodeStringshipping postal code
shippingCityStringshipping city
shippingStateStringshipping state
shippingCountryStringshipping country (3-alpha)
cvvVerifyResultStringM-- CVV2 Match
N-- CVV2 No Match
P--Not Processed
S--Issuer indicates that CVV2 data should be present on the card, but the merchant has indicated data is not present on the card
U--Issuer has not certified for CVV2 or Issuer has not provided Visa with the CVV2 encryption keys Empty--Transaction failed because wrong CVV2 number was entered or no CVV2 number was entered
paymentInstrumentIDStringID of payment Instrument: e.g. credit card number (hashed or encrypted) e.g. paypal account Id
paymentBillingAddressStringStreet Address , hashed or encrypted
paymentBillingPostalCodeStringpayment billing postal code
paymentBillingStateStringpayment billing state
paymentBillingCountryCodeStringpayment billing country code
paymentBillingNameStringName, hashed or encrypted, needs to be consistent with other names
isProxyIPStringWhether the IP address is a proxy or not
browserTypeStringI -- IE
C -- Chrome
F -- Firefox
O -- Other
browserLanguageStringSimilar to country code
paymentInstrumentTypeStringType of payments:
C -- Credit Card
D -- Debit Card
P -- Paypal
K -- Check
H -- Cash
O -- Other
cardTypeStringType of cards M -- Magnetic C -- Chip
cardNumberInputMethodStringInput method of payment instrument number:
K -- Keyed
S -- Swiped
C --- Chip
D -- Contactless
transactionDeviceTypeStringP -- PC
M -- Mobile Devices
C -- Console (e.g. Xbox, DVD)
O -- Other
transactionDeviceIdStringMac Address, or Hardware ID like serial number
transactionIPaddressStringFull IP Address for IPv4: 000.000.000.000
ipStateStringState of IP address originated from 2 alphabet letters
ipPostcodeStringPostal Code of IP address originated from
ipCountryCodeStringCountry code of IP address originated from
transactionDateStringDate when transaction occured Typically in the time zone of the processor, Format: yyyymmdd, e.g., 20000101
transactionTimeStringTime when transaction occurred. Typically in the time zone of processing end. Format: hhmmss, eg. 153059
localHourIntegerThe hour in local time. Value of 0-23
transactionScenarioStringA -- Authorization O -- Others
transactionTypeStringType of tranacation:
P -- Purchase
R -- Refund
T -- Transfer
O -- Other
transactionMethodStringI -- Internet (Online) Order
P -- Phone order
M -- Mail order
O -- Other

Account_Info.csv contain the following fields:

ColumnsTypeDescription
transactionDateStringDate when transaction occured Typically in the time zone of the processor. Format: yyyymmdd, e.g., 20000101
transactionTimeStringTime when transaction occurred. Typically in the time zone of processing end. Format: hhmmss, eg. 153059
accountOwnerNameStringUser name (hashed/encrypted)
accountAddressStringUser street address
accountPostalCodeStringUser postal code
paymentInstrumentAgeInAccountDoubleAge of payment instrument in the account
numPaymentRejects1dPerUserIntegerNumber of payment rejection in one day of this user
accountCityStringUser city
accountStateStringUser state
accountCountryStringUser country (3-alpha)
accountOpenDateStringAccount open date. Format: yyyymmdd
accountAgeIntegerAge of user account in number of days
isUserRegisteredStringWhether the user is registered or not

Fraud_Transactions.csv contain the following fields:

ColumnsTypeDescription
transactionIDStringUnique transaction Id
accountIDStringUnique account Id
transactionAmountDoubleTransaction amount in currency expressed in transactionCurrencyCode e.g., 12345.00
transactionCurrencyCodeStringCurrency code of the transaction. 3 alphabet letters, e.g., USD
transactionDateStringDate when transaction occured Typically in the time zone of the processor. Format: yyyymmdd, e.g., 20000101
transactionTimeStringTime when transaction occurred. Typically in the time zone of processing end. Format: hhmmss, eg. 153059
localHourIntegerThe hour in local time. Value of 0-23
transactionDeviceIdStringMac Address, or Hardware ID like serial number
transactionIPaddressStringFull IP Address for IPv4: 000.000.000.000