📊 Portfolio Optimization & Forecasting System

An advanced Python-based portfolio optimization and forecasting system that combines financial engineering, machine learning, and data science techniques to provide interactive portfolio analysis and optimization capabilities.

🌟 Key Features

• Multi-frequency Data Acquisition: Support for daily, weekly, monthly, and yearly stock data retrieval
• Comprehensive Financial Metrics: Calculation of key indicators including annualized return, volatility, Sharpe ratio, Sortino ratio, maximum drawdown, VaR, and CVaR
• Portfolio Optimization: Mean-variance optimization with multiple risk measures (MV, CVaR, MAD, etc.)
• Price Forecasting: Future stock price prediction using Prophet time series models
• Interactive Web Interface: Streamlit-based web application for user-friendly interaction
• Three-Portfolio Comparison: Historical-based, forecast-based, and actual realized portfolio optimization comparison
• Efficient Frontier Visualization: Graphical representation of optimal risk-return portfolios

🏗️ System Architecture

portfolio-optimization/
├── app.py                 # Streamlit web application
├── optimizer.py           # Portfolio optimization core logic
├── financial_indicator.py # Financial metrics calculation
├── get_data.py           # Data acquisition from Yahoo Finance
├── requirements.txt      # Project dependencies
└── data/                 # Data storage directory

📦 Installation

Prerequisites

• Python 3.8+
• pip package manager

Step-by-Step Installation

1. Clone or download the project files

git clone <https://github.com/Cloris-la/portfolio-forecasting.git>
cd portfolio-optimization

2. Create a virtual environment (recommended)

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies

pip install -r requirements.txt

4. Install additional dependencies if needed

pip install yfinance riskfolio-lib plotly streamlit prophet scipy

🚀 Quick Start

1. Run the Streamlit application

streamlit run src/app.py

2. Access the web interface

   • Open your browser and navigate to http://localhost:8501
   • Configure your analysis parameters in the sidebar
   • Click "Run Analysis" to start the optimization process

3. Explore different tabs

   • Data Overview: View historical price data and summary statistics
   • Financial Metrics: Analyze individual asset performance metrics
   • Portfolio Optimization: View optimal portfolio weights and expected performance
   • Performance Comparison: Compare historical, forecast, and actual portfolio performance
   • Efficient Frontier: Visualize the optimal risk-return tradeoff

🛠️ Usage Examples

Basic Portfolio Optimization

from optimizer import Portfolio_Optimizer
import pandas as pd

# Load your price data
prices = pd.read_csv('data/daily_stock_prices.csv', index_col=0, parse_dates=True)

# Initialize optimizer
optimizer = Portfolio_Optimizer(prices)

# Optimize for maximum Sharpe ratio
result = optimizer.optimize_portfolio(objective='Sharpe', risk_measure='MV')

print("Optimal weights:", result['weights'])
print("Expected return:", result['return'])
print("Volatility:", result['volatility'])

Financial Metrics Analysis

from financial_indicator import FinancialMetrics

metrics_calc = FinancialMetrics(frequency='daily')
metrics = metrics_calc.calculate_all_metrics(prices['AAPL'], 'AAPL')
print(metrics)

🔧 Configuration Options

Data Parameters

• Stocks: Select from 10+ popular stocks (AAPL, MSFT, GOOGL, NVDA, TSLA, NFLX, JPM, AMZN, META, DIS.)
• Date Range: Customizable start and end dates
• Frequency: Daily, weekly, or monthly data

Optimization Settings

• Risk Measures: MV (Mean-Variance), CVaR, MAD
• Objectives: Sharpe ratio maximization, risk minimization, return maximization
• Forecast Period: 10-90 days prediction horizon

📊 Output Features

1. Optimal Portfolio Weights: Percentage allocation for each asset
2. Performance Metrics: Expected return, volatility, Sharpe ratio, Sortino ratio
3. Visualizations:
   • Interactive price charts
   • Risk-return scatter plots
   • Portfolio composition pie charts
   • Efficient frontier graphs
4. Comparison Analysis: Side-by-side comparison of different optimization approaches

🎯 Use Cases

• Individual Investors: Personal portfolio optimization and risk management
• Financial Analysts: Rapid prototyping of investment strategies
• Educational Purposes: Learning modern portfolio theory and optimization techniques
• Research: Experimental analysis of different risk measures and optimization objectives

⚙️ Technical Details

Core Technologies

• Data Acquisition: yfinance for Yahoo Finance data
• Optimization Engine: Riskfolio-Lib for portfolio optimization
• Forecasting: Facebook Prophet for time series prediction
• Web Interface: Streamlit for interactive visualization
• Data Processing: Pandas, NumPy for financial calculations
• Visualization: Plotly for interactive charts

Mathematical Foundations

• Modern Portfolio Theory (Markowitz)
• Mean-Variance Optimization
• Conditional Value at Risk (CVaR)
• Sharpe and Sortino Ratios
• Time Series Forecasting

🐛 Troubleshooting

Common Issues

1. Optimization fails

   • Ensure you have sufficient historical data (>30 days)
   • Try different risk measures or objectives
   • Check for missing data or NaN values

2. Prophet installation issues

   • On Windows, you may need to install Microsoft C++ Build Tools
   • Use conda instead of pip: conda install -c conda-forge prophet

3. Data fetching errors

   • Check internet connection
   • Verify stock symbols are valid

Performance Tips

• Use weekly data for faster analysis with large date ranges
• Limit the number of stocks for quicker optimization
• Consider using simpler risk measures (MV) for faster computation

🤝 Contributing

We welcome contributions to improve this project! Areas for contribution include:

• Additional risk measures and optimization objectives
• Enhanced forecasting models
• Improved visualization capabilities
• Performance optimizations
• Documentation improvements

Please follow standard GitHub pull request procedures for contributions.

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

🔮 Future Enhancements

Planned features for future releases:

• Additional forecasting models (LSTM, ARIMA)
• Black-Litterman model integration
• Transaction cost consideration
• Tax optimization features
• Real-time data streaming
• Advanced risk analytics
• Portfolio backtesting capabilities
• API integration with brokerage accounts

---

Disclaimer: This tool is for educational and research purposes only. It does not provide financial advice. Always consult with a qualified financial advisor before making investment decisions. Past performance is not indicative of future results.