Multi-Asset Portfolio Optimizer

Project Overview

• This project implements a multi-asset portfolio optimization model in Python using historical financial data.
• It builds an efficient, diversified portfolio by applying mean-variance optimization (Markowitz Portfolio Theory) under practical constraints and analyzes its performance through backtesting and risk metrics.

---

Objectives

• Download and clean historical data for a diversified set of ETFs and asset classes.
• Calculate daily and monthly log returns, visualize trends, distributions, and correlations.
• Optimize portfolio weights to minimize risk for a target return, with constraints:
  • Fully invested portfolio
  • No short selling
  • Max 20% allocation per asset
• Generate the efficient frontier to analyze risk-return trade-offs.
• Backtest the optimized portfolio against:
  • An equal-weight portfolio
  • An SPY benchmark (S&P 500)
• Calculate performance metrics:
  • Annualized Return
  • Annualized Volatility
  • Sharpe Ratio
  • Maximum Drawdown
• Extend with turnover constraints to model practical rebalancing limits.

---

Technology Used

Python:

• Pandas
• NumPy
• matplotlib
• seaborn
• yfinance
• cvxpy

Assets Used

• VEA -> Developed Markets ex-US
• VWO -> Emerging Markets
• EWJ -> Japan
• TLT -> 20+ Year Treasury Bonds
• LQD -> Investment Grade Corporate Bonds
• GLD -> Gold
• VNQ -> REITs
• DBC -> Commodities Index

---

Key Steps

1. Data Acquisition: Download daily closing prices from 2015-2025 uding yfinance.
2. Data Cleaning and Preprocessing: Handle missing values, calculate daily and monthly log returns.
3. Exploratory Data Analysis: Plot -> Price Trends, Return Distributions, Correlation Heatmap.
4. Mean-Variance Optimization: Formulate the objective, set the constraints, solve using cvxpy quadratic programming.
5. Efficient Frontier: Calculate portfolios for varying target returns, visualize risk-return trade-offs.
6. Backtesting: Optimized portfolio VS equal-weight portfolio VS SPY benchmark
7. Extensions: Turnover constraints to limit drastic rebalancing.
8. Display results: View Reports folder.

---

Future Extensions

• Black-Litterman model for incorporating market views
• Robust optimization under parameter uncertainty
• Transaction cost modeling
• Interactive dashboard using Streamlit

---

Why this Project is Important:

• Portfolio optimization is a core concept in quantitative finance and asset management.
• This project demonstrates the ability to:
  • Integrate financial mathematics, statistics, and optimization
  • Build a full data pipeline from acquisition to analysis and evaluation
  • Apply convex optimization for real-world problems
  • Analyze risk-reward trade-offs for informed investment decisions

---

Author

Areeb Arshad

Sophomore, Data Science, Economics

Virginia Tech

---

Folder Structure

/
├── data/
│   ├── multiasset_benchmark.csv             
│   ├── multiasset_closing_prices.csv            
│   ├── multiasset_cum_returns.csv
|   ├── multiasset_daily_returns.csv
|   ├── multiasset_ewc.csv
|   ├── multiasset_stats.csv
|   ├── README.md                       
├── notebooks/
│   ├── Backtesting/      
│   ├── DataAcquisition_and_Preprocessing/            
│   ├── Extensions_Dashboard/
│   ├── Final_Model/
|   ├── Portfolio_Optimization/
|   ├── README.md     
├── plots/
│   ├── Closing_Prices/               
│   ├── Cumulative_Returns/ 
│   ├── Daily_Return_Distributions/       
│   ├── Efficient_Frontier/                     
│   ├── Optimal_Weights/               
│   ├── Returns_Correlation_Heatmap/
|   ├── README.md           
├── reports/
|   ├── README.md            
└── README.md