Dark Mode
European Bank Customer Retention Data
Finance & Banking Analytics
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
Data explains the factors influencing customer churn in the banking domain, serving as a foundation for binary classification tasks. This synthetic dataset was developed for a Playground Series challenge and includes diverse features, alongside various engineered metrics, to capture relevant information about customer demographics and financial behaviour. It provides analysts with the metrics needed to build predictive models and gain insights into customer retention dynamics.
Columns
The structure contains 25 distinct columns, including original customer variables and derived engineered metrics:
- Exited: The primary target variable, indicating whether a customer has churned (1) or not (0).
- Surname: Label-encoded representation of the customer's surname, also used to derive Surname TFIDF features (0-4).
- CreditScore: A numerical measure of creditworthiness, ranging from 350 to 850 (mean 656).
- Age: Customer age, ranging from 18 to 92 (mean 38.2).
- Tenure: The length of time a customer has held an account, ranging from 0 to 10 years (mean 5.02).
- Balance: The account balance held, with a maximum value of 251k and a mean of 56.7k.
- NumOfProducts: The count of products used by the customer (range 1 to 4).
- HasCrCard: A binary feature indicating the presence of a credit card.
- IsActiveMember: A binary feature representing active membership status.
- EstimatedSalary: The customer's estimated salary, ranging up to 200k (mean 112k).
- Geography (France, Spain, Germany): One-hot encoded features indicating the country of residence.
- Gender (Male, Female): One-hot encoded gender representation.
- Mem__no__Products: An engineered metric derived from the product of the number of products and active membership status.
- Cred_Bal_Sal: A relative measure of financial health, calculated as the ratio of the product of credit score and balance to estimated salary.
- Bal_sal: The balance-to-salary ratio.
- Tenure_Age: The tenure-to-age ratio.
- Age_Tenure_product: A feature capturing the interaction between age and tenure.
Distribution
The data file is usually provided in CSV format. The structure includes 175k valid records across all fields. Data validity is 100%, with no missing values reported for core fields. The mean value of the target variable 'Exited' is 0.21, suggesting approximately 21% of customers have churned. The expected update frequency for this resource is annually.
Usage
This synthetic banking dataset is suitable for several analytical purposes:
- Predictive Modelling: Utilising machine learning techniques to classify customers as likely to exit or remain with the bank.
- Binary Classification: Serving as input for binary classification tasks related to customer attrition.
- Insight Generation: Providing features that allow exploration of key indicators influencing customer retention dynamics.
- Data Visualization: Supporting advanced data storytelling and plotting related to financial behavior.
Coverage
The geographical scope covers customers residing in France, Spain, and Germany. Demographic coverage includes age, ranging from 18 to 92, and gender (male and female). Financial coverage includes detailed metrics such as credit score (350 to 850), account balance, and estimated salary.
License
CC0: Public Domain
Who Can Use It
This dataset is intended for:
- Data Scientists: For training and validating predictive models designed to assess customer risk and retention.
- Analysts: To gain actionable insights into customer banking behavior and financial health metrics.
Dataset Name Suggestions
- Bank Customer Churn Predictor
- Financial Attrition Binary Classification
- European Bank Customer Retention Data
- Synthetic Bank Churn Analysis
Attributes
Original Data Source: European Bank Customer Retention Data
QUALITYLoading...