How ayushbaral/s_score works
Project type and purpose
• This is a data-analysis/algorithmic-trading toolkit implemented as a pair of Jupyter notebooks.
• Its purpose is to download 1-year historical price data for 508 stocks + SPY, compute an S-Score via a PCA-based statistical arbitrage model (Avellaneda & Lee), and emit buy/sell signals.
File structure
• Top-level files:
• download_data.ipynb – data ingestion notebook
• signals.ipynb – core algorithm and signal generation notebook
• 500_stocks_1yr.csv, SPY_1yr.csv – raw data outputs
• requirements.txt – Python deps (pandas, numpy, scikit-learn, statsmodels, etc.)
• README.md – project overview & algorithm steps
• No frontend/backend split; it’s a linear data pipeline orchestrated in notebooks.
README for additional insights
• Algorithm stages enumerated: return computation → standardization → covariance matrix → PCA → eigenportfolio returns → regression residuals → OU process test & AR(1) → S-Score → trading signals.
Detail without over-splitting
• Dependencies (e.g., PCA module depends on covariance matrix).
• Environment/dependency management is shown (virtual environment → requirements.txt → runs notebooks).