Introduction

This notebook was made as part of the final project of Data Science at Rakamin Academy

Problem Statement

Customer churn is a measure of the number of customers who have stopped using the company's products or services. To measure the customer churn ratio can be calculated by dividing the number of customers lost by the number of customers acquired (Accurate, 2021). Based on this data, the customer churn rate at the Bank reached 20.38%.

Goals and Objective

Identify the factors that led customers to churn and using machine learning models to predict the probability of customer churn.

Dataset Information

The dataset from (Kaggle) contains information from all customers who were churn from the Bank. This data contains 14 columns and 10000 rows where each column includes:

• RowNumber : Number listed on each row
• CustomerId : ID number for each customer
• Surname : Family/last name for each customer
• CreditScore : Number of credit scores owned by customers
• Geography : Location of customer addresses by country
• Gender : Customer gender information (Male/Female)
• Age : Information on the age of each customer
• Tenure : Number of customer tenors at Bank Java by year
• HasCrCard : Customers who have a Bank Java credit card (0 = No, 1 = Yes)
• EstimatedSalary : Estimated salary value of each customer
• IsActiveMember : Customer who is active or not (0 = Inactive, 1 = Active)
• Exited : Customer who are churn or not (0 = Not Churn, 1 = Churn)