Part of Applied Data Science with Machine Learning Bootcamp
Customer acquisition is usually more costly to the business than retaining an existing customer. In this EDA we look at churn data for credit card customers, to analyze the possible factors involved in churn, for the bank to have actionable insights.
EDA will not be using ML techniques.
Data set is acquired from https://www.kaggle.com/sakshigoyal7/credit-card-customers
It contains information on 10,000 customers, with 18 features for analysis. Attrition in this data set is at 16.07%.
| Attribute | Definition |
|---|---|
| CLIENTNUM | Unique identifier for the client |
| Attrition_Flag | Attrited or Existing |
| Customer_Age | Customer's age in years |
| Gender | M/F |
| Dependent_Count | Number of dependents |
| Education_Level | Qualification level of account holder |
| Marital_Status | Married, Single, Divorced, or Unknown |
| Income_Category | Income category in 4 brackets |
| Card_Category | Product variable: Blue, Silver, Gold, Platinum |
| Months_on_book | Period of relationship with the bank |
| Total_Relationship_Count | Total no. of products held by the customer |
| Months_Inactive_12_mon | No. of mths inactive in the past 12 mths |
| Contacts_Count_12_mon | No. of contacts with the bank in the past 12 mths |
| Credit_Limit | Credit limit on the card |
| Total_Revolving_Bal | Revolving balance is credit that is carried over, high revolving balance indicates a reliance on credit |
| Avg_Open_To_Buy | Avg open to buy credit line |
| Total_Amt_Chng_Q4_Q1 | Change in txn amt Q4 over Q1 |
| Total_Trans_Amt | Total txn amt (last 12 mths) |
| Total_Trans_Ct | Total txn count (last 12 mths) |
| Total_Ct_Chng_Q4_Q1 | Change in txn count Q4 over Q1 |
| Avg_Utilization_Ratio | Avg card utilization ratio |