Lloyds Banking Group Data Challenge

Lloyds Banking Group is launching a new loans product. Prior to the launch they would like you to use historical customer data to:

• Task 1 – Data Strategy : Understand and summarise the different behaviours or attributes between customers who paid back their loan and customers who did not
• Task 2 – Data Science: To use the historical data to design a process which predicts the likelihood of a new customer not paying back their loan

1. Understanding the data

2. Exloratory Data Analysis

2.1 Annual Income

2.2 Public bankruptcies

2.3 Loan Amount

2.4 Affordability

2.5 Show the pair plot

3. Using machine learning model to predic the default.

3.1 Logistic Regression model - 0.79

3.2 Ramdom Forest model - 0.79

3.3 XGBoost model - 0.78

4. Compare Gini performance for both algorithms

• Number of records: 18,324

Result