📈 Pairs Trading - Quantitative Research Project

🔎 Objective

This project implements a Statistical Arbitrage - Pairs Trading Strategy analyzing 20 US Tech Stocks from the S&P 500, using cointegration, mean-reversion, and backtesting techniques.

✅ Key steps:

• Fetching historical stock data (2005-2025)
• Testing pairs for cointegration
• Backtesting the strategy
• Generating trading signals
• Visualizing performance

---

📅 Dataset Details

• Stocks: AAPL, MSFT, GOOG, AMZN, META, NVDA, TSLA, NFLX, CRM, IBM, PYPL, etc.
• Source: yfinance API
• Data Range: 2005 - 2025
• Frequency: Daily (Adjusted Close Prices)

---

⚙️ Project Pipeline

1. Data Fetching (fetch_data.py)

   • Pulled tech stock data for 20 years (2005-2025)
   • Saved in data/tech_stocks_data.csv

2. Cointegration Testing (check_cointegration.py)

   • Ran Engle-Granger cointegration tests
   • Filtered pairs with significant p-values (p < 0.05)

3. Backtesting (backtest.py)

   • Created the spread of cointegrated pairs
   • Checked the spread behavior over time
   • Calculated PnL for basic mean-reversion strategy

4. Trading Logic (trading_logic.py)

   • Applied 1 standard deviation threshold rules
   • Generated buy/sell signals

5. Visualization

   • Plotted spreads and trading signals
   • Saved plots in images/

---

🗂 Project Structure

``

---

📊 Key Learnings

✅ Learned how cointegration forms the basis for Pairs Trading
✅ Understood mean-reversion concept and backtesting strategies
✅ Hands-on data extraction, statistical testing, and signal generation
✅ Visualized results for better analysis and reporting

---

💻 Technologies Used

• Python 3
• Libraries: pandas, numpy, matplotlib, statsmodels, yfinance
• Jupyter Notebook
• Git / GitHub

---

📈 Example Visualizations (Saved in /images/)

• Spread plots of pairs over time
• Trading signal plots with entry/exit points

---

✅ Next Steps / Improvements

• Test with different sectors (Finance, Energy, Healthcare)
• Explore dynamic hedge ratios (Kalman Filter / PCA)
• Include transaction costs and slippage
• Automate the full pipeline

---

📬 Project Info

• Author: Niraj Patil
• Project Duration: 3-4 Days
• Timeframe of Data: 2005 - 2025
• GitHub Repo: Pairs-Trading-Quant-Research

---

⭐ Conclusion

This project helped me build a strong foundation in Statistical Arbitrage - Pairs Trading, covering data handling, cointegration tests, backtesting logic, and visualization.

It's the first of 10 Quant Finance projects on my roadmap to strengthen my Quant profile and secure a role in top quantitative finance firms.