Card fraud

Dataset Features:

• Transaction_ID: Unique identifier for each transaction.
• User_ID: Unique identifier for each user involved in a transaction.
• Transaction_Amount: The monetary amount of the transaction.
• Transaction_Date: Date of the transaction.
• Transaction_Time: Time of the transaction.
• Transaction_Location: Geographic location where the transaction occurred.
• Merchant_ID: Unique identifier for the merchant involved in the transaction.
• Device_ID: Identifier of the device used for the transaction.
• Card_Type: Type of card used (e.g., UzCard, Humo).
• Transaction_Currency: Currency in which the transaction was processed.
• Transaction_Status: Status of the transaction (e.g., Successful, Failed, Reversed).
• Previous_Transaction_Count: Number of prior transactions made by the user.
• Distance_Between_Transactions_km: Geographical distance between this transaction and the previous one.
• Time_Since_Last_Transaction_min: Time elapsed since the last transaction by the user.
• Authentication_Method: Method used to authenticate the transaction (e.g., 2FA, Biometric).
• Transaction_Velocity: Rate of transactions by the user in a given period.
• Transaction_Category: Type of transaction (e.g., Cash In, Cash Out, Payment).
• isFraud: Binary indicator of whether the transaction was fraudulent (1 for fraud, 0 for non-fraud).

Distribution

• Data Format: Tabular data (CSV).
• Number of Records: 100000 rows
• Number of Columns: 18 columns, including identifiers, attributes, and outcomes.
• Column Types:
  • Numeric Columns: Transaction Amount, Previous Transaction Count, Distance Between Transactions, Time Since Last Transaction, Transaction Velocity.
  • Categorical Columns: User ID, Transaction Location, Merchant ID, Device ID, Card Type, Transaction Currency, Authentication Method, Transaction Category, Transaction Status.
  • Datetime Columns: Transaction Date, Transaction Time.
  • Binary Column: isFraud.
• File Size: 11.59 MB
• Dataset Structure: Each row represents a unique transaction, and each column provides a specific feature related to that transaction.

Usage

• Fraud Detection Models: Train machine learning algorithms to classify fraudulent transactions.
• Anomaly Detection: Identify unusual patterns in transaction behavior.
• Behavioral Analysis: Analyze user and merchant transaction patterns.
• Authentication Method Efficiency: Study the effectiveness of different authentication methods in preventing fraud.
• Risk Assessment: Develop tools for assessing the risk of fraud in financial transactions.

Coverage

• Geographic Coverage: Includes transaction data from various regions such as Surkhandarya, Namangan, and Navoiy, among others.
• Time Range: Data is collected starting from January 1, 2024, to January 7, 2024 (specific end date depends on the full dataset).
• Demographics : Covers a range of user types, transaction categories, and merchant activities.

License

Who Can Use It

• Data Scientists: For training and validating machine learning models.
• Researchers: For academic or scientific studies on fraud detection and prevention.
• Businesses: For enhancing fraud monitoring systems and developing insights into transaction patterns.
• Students and Educators: for teaching and learning data analysis, machine learning, and cybersecurity principles.