Financial Time Series Analysis with Python

English | 한국어

A comprehensive open-source project for mastering financial time series analysis, algorithmic trading strategies, and production-level automated trading systems using Python. This project combines theoretical knowledge with practical implementation, featuring a live production trading system that is currently operating with real capital.

🎓 Udemy Course

Mastering Financial Time Series Analysis with Python

👉 Enroll in the Course

• All Sections: Course materials are fully uploaded and available
• Section 3: Live production system with real-time trading records

📊 Live Trading Records

Currently in Live Production

This project includes a production trading system that is actively trading with real capital. Daily trading records are automatically uploaded to:

🌐 Trading History Dashboard

The system uses broker API integration for automated trading, and all trades are transparently displayed on the website.

📚 Project Structure

Section 1: Financial Time Series Analysis

Status: ✅ Course Available

Comprehensive coverage of time series fundamentals and advanced techniques:

• Chapter 1: Fundamentals of Time Series Data Analysis

  • Stationarity and non-stationarity
  • Differencing and transformation
  • Seasonal decomposition

• Chapter 2: Advanced Time Series Analysis

  • ADF (Augmented Dickey-Fuller) test
  • AR (Autoregressive) models
  • PACF (Partial Autocorrelation Function) analysis
  • Random walk theory

• Chapter 3: Univariate Time Series Analysis

  • AR, MA, ARMA models
  • AIC vs BIC model selection
  • Auto-ARIMA
  • Ljung-Box test for residual analysis

• Chapter 4: Advanced Volatility Modeling and Forecasting

  • ARCH models
  • GARCH models
  • ARIMA-GARCH hybrid models
  • Backtesting strategies

• Chapter 5: Multivariate Time Series Analysis

  • VAR (Vector Autoregression) models
  • VARMA models
  • Granger causality analysis

• Chapter 6: Advanced Multivariate Time Series Analysis

  • VECM (Vector Error Correction Model)
  • Johansen cointegration test
  • VAR IRF (Impulse Response Function)
  • VAR FEVD (Forecast Error Variance Decomposition)
  • VECM-APARCH hybrid models

Section 2: Advanced Investment Strategy Design

Status: ✅ Course Available

Practical implementation of trading strategies:

• Chapter 1: Dynamic Time Series Simulations

  • VECM-EGARCH hybrid model
  • Dynamic re-optimization
  • Long/short position management

• Chapter 2: Applying Strategies to Bitcoin Trading

  • Bitcoin-specific optimizations
  • Commission fee considerations
  • Volatility-based re-optimization

• Chapter 3: AI Trading Using Binance

  • Binance API integration
  • Real-time signal generation
  • Automated order execution

Section 3: Production Investment Strategy

Status: ✅ Course Available 🚀 Live Production System

A production-level trading system currently operating with real capital:

• Chapter 1: VECM-EGARCH Hybrid Model

  • Vector Error Correction Model for cointegration relationships
  • Exponential GARCH for volatility modeling
  • Dynamic position sizing based on model confidence
  • Key Features:
    • ✅ Information leakage prevention (walking forward validation)
    • ✅ Dynamic re-optimization based on ECT alpha
    • ✅ Confidence-based position sizing (0.2~0.8 fraction range)
    • ✅ Separate forecast horizons for buy (4 days) and sell (7 days)

• Chapter 2: Reinforcement Learning (RL) Strategy

  • VECM-GARCH Hybrid combined with RL Agent
  • Market Regime Detection (Bull, Bear, Sideways, High Vol)
  • Simple Policy RL agent for dynamic position blending
  • Adaptive confidence thresholds

• Mathematical Models:

  VECM Model:

  ΔY_t = αβ'Y_{t-1} + Γ₁ΔY_{t-1} + ... + Γ_{p-1}ΔY_{t-p+1} + ε_t
  

  • α: adjustment coefficients (speed of adjustment to equilibrium)
  • β: cointegration vectors (long-run relationships)
  • Γᵢ: short-run dynamics coefficients

  EGARCH Model:

  log(σ²_t) = ω + Σᵢ₌₁ᵖ (αᵢ|z_{t-i}| + γᵢz_{t-i}) + Σⱼ₌₁ᵠ βⱼlog(σ²_{t-j})
  

  • Captures asymmetric volatility effects
  • Ensures positive variance through log transformation

  Hybrid Forecast:

  Ŷ_{t+1} = VECM_forecast + EGARCH_mean_adjustment
  

• Trading Strategy:

  • Long Entry: hybrid_yhat_buy > actual_price AND lower_price < lower_bound_mean
  • Long Exit: upper_price > upper_bound_mean
  • Dynamic Re-optimization: When ECT alpha changes from negative to positive

Section 4: Advanced Time Series Models

Status: ✅ Course Available

📖 Detailed Documentation

Modern statistical and ML techniques for challenging temporal structures:

• Chapter 1: State-Space Models – time-varying beta tracking and error-correction diagnostics
• Chapter 2: Kalman Filter suite (Custom, FilterPy, PyKalman EM, Particle) – dynamic beta tracking with quantile-based switching
• Chapter 3: Prophet Model – seasonality-aware forecasting with rolling re-training
• Chapter 4: Deep Learning (LSTM) – direction classification with imbalance-aware training
• Chapter 5: Tree-Based ML (XGBoost) – binary classification for direction prediction with rich technical features and ROC-optimized thresholds
• Chapter 6: Wavelet Transform – multi-resolution feature engineering for volatility regimes
• Chapter 7: Copula Models – dependence modeling and tail-risk simulation

Section 5: Factor-Based Asset Pricing Models

Status: ✅ Course Available

📖 Detailed Documentation

Theoretical foundations and practical applications of factor-based asset pricing models:

• Chapter 1: CAPM Limitations and Fama-French Model Origins

  • Empirical testing of CAPM
  • Identifying market anomalies (Size, Value)
  • Visualizing model limitations

• Chapter 2: Fama-French 3-Factor Model

  • Implementing the 3-factor model
  • Calculating factor exposures (Betas)
  • Comparing multi-factor models vs CAPM

• Chapter 3: Fama-French 5-Factor and Extended Models

  • Profitability (RMW) and Investment (CMA) factors
  • Momentum factor integration (6-Factor model)
  • Model selection and comparison

• Chapter 4: Practical Application and Backtesting

  • Factor-based portfolio construction
  • Walk-forward validation
  • Transaction cost analysis
  • Performance evaluation (Sharpe, Alpha, etc.)

Appendix: Financial Mathematics Theory and Practical Examples

Status: ✅ Available

📖 Detailed Documentation

A comprehensive guide to all financial mathematics theory used in quant trading, implemented with easy-to-understand example code:

• Chapter 1: Linear Algebra

  • Portfolio optimization
  • PCA-based factor analysis
  • Multi-factor regression (Fama-French)
  • Matrix operations in VAR & VECM models

• Chapter 2: Analysis & Calculus

  • Gradient descent visualization
  • Understanding backpropagation algorithm
  • Calculus principles in GARCH models
  • Wavelet Transform
  • Ito's Lemma
  • Bayesian Optimization

• Chapter 3: Probability & Time Series Statistics

  • Stationarity testing and understanding
  • Probabilistic foundations of ARIMA models
  • Cointegration and pair trading
  • Dependence analysis using Copula
  • Monte Carlo simulation

• Chapter 4: Bayesian Statistics & Filtering

  • Bayesian inference examples
  • Understanding Kalman Filter
  • State-space models

🚀 Quick Start

Installation

1. Clone the repository:

git clone https://github.com/leesh2015/financial-timeseries-python.git
cd financial-timeseries-python

2. Install dependencies:

pip install -r requirements.txt

Running Examples

Section 1 - Time Series Analysis:

cd "Section1.Financial Time Series Analysis/Chapter1.Fundamentals of Time Series Data Analysis"
python stable_data.py

Section 2 - Strategy Design:

cd "Section2.Advanced Investment Strategy Design/Chapter1.Dynamic Time Series Simulations"
python dynamic_simulation.py

Section 3 - Production Simulation:

# Chapter 1: VECM-EGARCH Hybrid
cd "Section3.Production Investment Strategy/Chapter1.VECM-EGARCH Hybrid"
python production_simulation_.py

# Chapter 2: Reinforcement Learning
cd "Section3.Production Investment Strategy/Chapter2.Reinforcement Learning"
python dynamic_simulation_rl.py

Section 4 - Advanced Time Series Models:

cd "Section4.Advanced Time Series Models/Chapter1.State-Space Models"
python state_space_model.py

Section 5 - Factor Models:

cd "Section5.Factor-Based Asset Pricing Models/Chapter4.Practical Application and Backtesting"
python factor_portfolio_backtest.py

Appendix - Financial Mathematics:

# Install dependencies from project root (skip if already installed)
pip install -r requirements.txt

cd Appendix

# Chapter 1: Linear Algebra
python Chapter1_Linear_Algebra/portfolio_optimization.py

# Chapter 2: Calculus
python Chapter2_Calculus/gradient_descent_demo.py

# Chapter 3: Probability & Statistics
python Chapter3_Probability_Statistics/stationarity_analysis.py

# Chapter 4: Bayesian
python Chapter4_Bayesian_Filtering/kalman_filter_demo.py

Results will be saved in the results/ folder within each section.

📦 Dependencies

Core dependencies (see requirements.txt for full list):

• Data Science: numpy, pandas, scipy
• Time Series: statsmodels, arch, pmdarima
• Data Collection: yfinance
• Visualization: matplotlib, seaborn
• Machine Learning: scikit-learn
• Cryptocurrency: ccxt (for Binance integration)
• Excel Support: openpyxl

🎯 Key Features

Information Leakage Prevention

• Uses data up to time t-1 to predict price at time t
• Models are re-trained at each step using only historical data (walking forward)
• No future data is used in any prediction or optimization step

Dynamic Model Adaptation

• Automatic re-optimization when market conditions change
• ECT alpha monitoring for cointegration relationship health
• Volatility-based model adjustments

Confidence-Based Position Sizing

• Dynamic position sizing based on VECM model confidence
• Fraction range: 0.2 (low confidence) to 0.8 (high confidence)
• Adaptive threshold calculation using rolling window

Production-Ready

• Real-time broker API integration
• Automated trade execution
• Daily performance tracking and reporting
• Transparent trade history on web dashboard

📈 Performance Metrics

The production system tracks comprehensive performance metrics:

• Total P&L
• Win Rate
• Sharpe Ratio
• Maximum Drawdown
• Annualized Returns
• Buy-and-Hold Comparison

View live metrics at: Trading History Dashboard

🔬 Research & Methodology

This project implements state-of-the-art financial econometrics techniques:

• Cointegration Analysis: Identifying long-run equilibrium relationships
• Error Correction Models: Capturing short-term deviations from equilibrium
• GARCH Family Models: Modeling volatility clustering and asymmetry
• Hybrid Forecasting: Combining multiple models for improved accuracy
• Dynamic Optimization: Adapting to changing market regimes

📖 Course-Code Mapping

Course Section	Repository Section	Status
Section 1: Time Series Fundamentals	Section1.Financial Time Series Analysis/	✅ Available
Section 2: Strategy Design	Section2.Advanced Investment Strategy Design/	✅ Available
Section 3: Production System	Section3.Production Investment Strategy/	✅ Available 🚀
Section 4: Advanced Time Series Models	Section4.Advanced Time Series Models/	✅ Available
Section 5: Factor Models	Section5.Factor-Based Asset Pricing Models/	✅ Available
Appendix: Financial Mathematics	Appendix/	✅ Available

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request. For major changes, please open an issue first to discuss what you would like to change.

Join the Discussion: Have questions, ideas, or want to share your results? Join our GitHub Discussions to connect with the community!

📄 License

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

⚠️ Disclaimer

Important: This project is for educational and research purposes. The production trading system is provided as a demonstration of the concepts taught in the course.

• Past performance does not guarantee future results
• Trading involves risk of financial loss
• Always conduct thorough backtesting before deploying any trading strategy
• The authors are not responsible for any financial losses incurred from using this code

🔗 Links

• Udemy Course: Mastering Financial Time Series Analysis with Python
• Trading Dashboard: leenaissance.com/trading-history
• GitHub Discussions: Join the Community

📧 Contact

For questions, suggestions, or collaboration opportunities, please open an issue on GitHub or contact through the Udemy course platform.

---

Made with ❤️ for the algorithmic trading community

This project demonstrates the complete journey from theoretical time series analysis to production-level automated trading systems.