Loan Analysis Project

Project Overview

Analysis of a large loan portfolio ($30.1B) from a major peer-to-peer lending platform that reveals significant opportunities for risk management improvement. The study identified critical risk patterns and developed data-driven solutions for enhancing loan decisioning.

Key Findings:

• Overall default rate: 13.05%
• High-risk grades (8.4% of portfolio) drive 31.1% of defaults
• 60-month terms show 60% higher defaults than 36-month terms
• Geographic variations exceed 15% default rates in highest-risk states

Installation

pip install -r requirements.txt

Project Structure

├── data/
│   ├── raw/       # Original loan data
│   ├── interim/   # Partially processed data
│   ├── cleaned/   # Cleaned dataset
│   └── processed/ # Fully processed data
├── notebooks/
│   ├── 1_data_cleaning.ipynb
│   ├── 2_feature_engineering.ipynb
│   ├── 3_exploratory_data_analysis.ipynb
│   └── 4_statistical_analysis.ipynb
├── src/
│   └── data/
│       └── make_dataset.py
└── docs/
    ├── technical.md
    ├── analysis.md
    └── implementation.md

Quick Start

1. Install dependencies:

   pip install -r requirements.txt

2. Run data pipeline:

   python src/data/make_dataset.py

3. Execute notebooks in sequence
4. View results in final notebook

Documentation

For detailed documentation, see:

• Technical Documentation - Data overview and technical implementation
• Analysis Results - Detailed findings and insights
• Implementation Guide - Solutions and next steps