Quantitative-Research

A collection of advanced quantitative finance and risk modeling tasks.

---

📚 Overview

This repository contains a suite of Python-based solutions for real-world quantitative research and risk management problems, including:

• Natural Gas Price Estimation & Storage Contract Valuation
• Loan Default Probability (PD) Modeling & Expected Loss Calculation
• FICO Score Quantization & Rating Map Optimization

Each module is designed for clarity, extensibility, and practical use in a professional or academic setting.

---

🚀 Project Structure

• gas_price_analysis.py
  Estimate natural gas prices for any date and value storage contracts with custom cash flows.
• loan_default_analysis.py
  Predict loan default probability (PD) and expected loss using machine learning.
• task4.py
  FICO score bucketing and quantization using dynamic programming and log-likelihood/MSE optimization.
• Task 3 and 4_Loan_Data.csv
  Loan borrower dataset for PD modeling and quantization.
• Nat_Gas.csv
  Monthly natural gas price data for time series modeling.
• requirements.txt
  All required Python dependencies.

---

🏆 Key Features

1. Natural Gas Price Estimation & Storage Contract Valuation

• Interpolate and extrapolate gas prices for any date (past or future)
• Value storage contracts with custom injection/withdrawal schedules, storage costs, and volume constraints
• Detailed cash flow breakdown and storage level tracking

2. Loan Default Probability & Expected Loss Modeling

• Train and compare Logistic Regression and Random Forest models
• Predict PD for any borrower profile
• Calculate expected loss using a configurable recovery rate
• Feature importance visualization for model interpretability

3. FICO Score Quantization & Rating Map

• Optimal bucketing of FICO scores using dynamic programming
• Supports both log-likelihood and MSE as optimization criteria
• Maps FICO scores to ratings for robust credit risk segmentation

---

📝 Example Outputs

🔹 Natural Gas Storage Contract

Contract Value: $-5,995.53

Cash Flows:
2023-01-01: injection    $-23,232.26
2023-02-01: storage_cost $-930.00
... (truncated)

🔹 Loan Default Analysis

Model Performance:
Logistic Regression ROC AUC: 1.0000
Random Forest ROC AUC: 0.9997

Example Loan Analysis:
Loan Amount: $10,000.00
Probability of Default: 2.13%
Expected Loss: $1,917.00

🔹 FICO Score Bucketing

Optimal FICO boundaries: [545, 602, 650, 700, 750]

Sample mapping:
   fico_score  rating
0         605       2
1         720       5
...

---

🛠️ How to Run

1. Install dependencies:

   pip install -r requirements.txt

2. Run any module:

   python gas_price_analysis.py
   python loan_default_analysis.py
   python task4.py

---

📖 References & Further Reading

• Quantization (signal processing)
• Likelihood function
• Dynamic programming

---

💡 Contributing

Pull requests and suggestions are welcome! For major changes, please open an issue first to discuss what you would like to change.

---

© 2025 RichieOnData