Skip to main content

churn_retail_output_final

CUSTOMER​

  • Description: Represents the customer number.
  • Example: 67890
  • Type: String

DEBIT_TXNS_PRCNT​

  • Description: The percentage change in debit transactions over the recent 6 months compared to the previous 6 months. For example, if April–October 2024 are the recent six months, the comparison would be against April–October 2023.
  • Formula: ((R6 - P6) / P6) * 100, where R6 = recent 6 months and P6 = previous 6 months.
  • Example: 15.67
  • Type: Decimal

CREDIT_TXNS_PRCNT​

  • Description: Same as DEBIT_TXNS_PRCNT, but for credit transactions.
  • Formula: ((R6 - P6) / P6) * 100
  • Example: -5.23
  • Type: Decimal

REVENUE_PRCNT​

  • Description: Similar to DEBIT_TXNS_PRCNT, but measures the percentage change in revenue generated by the customer for the bank over the recent 6 months compared to the previous 6 months.
  • Formula: ((R6 - P6) / P6) * 100
  • Example: 25.89
  • Type: Decimal

DEBIT_TXN_CHNGE​

  • Description: Describes whether debit transactions have increased or decreased.
  • Example: "Increased"
  • Type: String

CREDIT_TXN_CHNGE​

  • Description: Describes whether credit transactions have increased or decreased.
  • Example: "Decreased"
  • Type: String

REVENUE_CHNGE​

  • Description: Describes whether the customer's revenue has increased or decreased.
  • Example: "Increased"
  • Type: String

RECENCY​

  • Description: Indicates how recent the last transaction was, typically categorized as either greater than or less than 30 days.
  • Example: "Less than 30 days"
  • Type: String

PREDICTED​

  • Description: The churn model's prediction, where 1 indicates a high likelihood of churn (probability > 0.5), and 0 indicates a low likelihood of churn (probability ≤ 0.5).
  • Example: 1
  • Type: Integer

PROBABILITIES_0​

  • Description: The probability that the customer will not churn.
  • Example: 0.30
  • Type: Decimal

PROBABILITIES_1​

  • Description: The probability that the customer will churn.
  • Example: 0.70
  • Type: Decimal

MODEL_RUNTIME​

  • Description: The date the churn model was executed.
  • Example: 2024-10-01
  • Type: Date