Click here to download detailed EDA Report

Problem Statement

This project aims to analyze the loan application dataset by Home Credit Group and identify factors that significantly contribute to loan defaults. Specfically the project seeks to

• Understand the structure, content, and quality of the loan application dataset.
• Identify key demographic, financial, and application-related factors that are significantly associated with loan defaults.
• Examine the distribution of individual variables and investigate relationships between pairs of variables, focusing on those that correlate with loan defaults.
• Provide actionable recommendations to improve risk assessment and optimize the lending portfolio.

Approach

• Understanding the business requirments
• Import relevant libraries
• Load the dataset
• Data Preprocessing
  • Handling missing values
  • Outlier Treatment
• Exploratory Data Analysis
  • Univariate Analysis (Categorical)
  • Correlation Between Numerical Variables
  • Univariate Analysis (numerical)
  • Bivariate Analysis
  • Scope for Further Analysis
• Key Takeaways and recommendations

Exploratory Data Analysis

Univariate Analysis (Categorical)

• Share of each loan category and the default rates
• Loan default rates based on gender
• Loan default based on who accompanied the client
• Indepth analysis between client profile (income type, education, family status, housing type, occupation) and loan default
• Commercial clients: distribution, default percentages

Univariate Analysis (numerical)

• Analysis of credit amounts: median, distribution, anomalies
• Nature of relationship between client income and default rates
• Most common annuity amounts
• Distribution of Asset value in case of both default and non-default

Bivariate Analysis

• Realtionship between credit amount and asset valuation
• Safest Credit ranged based on client income and asset valuation
• Relation between client income and credit amount
• Variation in credit amount based on family strength

Key Takeaways and Recommendations

• Targeting:
  • Low-Risk Groups: Widows, married individuals, those with higher education, accountants,commercial clients with office apartments, and clients who own a house or apartment.
  • High-Risk Groups: Unemployed individuals, mothers on maternity leave, those with lower secondary education, people in civil marriages, single people, people who rent, low-skill laborers, clients who live with parents, clients from transport, industry types 3 and 8.
• Urgent policy adjustments needed for clients on maternity leave either through enhanced credit mechanisms or social security measures.
• Introducing new schemes and benefits to unemployed individuals aimed at upskilling them for employment and debt repayment
• More room to grow by increasing focus on commercial clients
• Among consumer clients those owning a house/apartment are the safest target group
• Safest Credit Range: Total income greater than $150,000 and Credit Amount between 2 million and 2.5 million