This project builds a data pipeline for selecting a tradable universe of US equities and computing a broad set of technical indicators for downstream analysis or modeling.
The focus is on data preparation, not prediction.
Most trading and ML projects skip an important step: how the stock universe and features are constructed.
This project addresses that gap by:
- Filtering stocks based on price constraints
- Maintaining a reusable stock universe
- Downloading historical OHLCV data
- Computing a wide range of technical indicators
- Producing clean, indicator-rich datasets
These outputs can later be used for:
- Strategy research
- Feature selection
- Signal evaluation
- Machine learning models
- Market Data: Yahoo Finance (yfinance)
- Indicators: stockstats
- Universe: US equity ticker lists
- Filter US stocks by price range
- Persist selected tickers
- Avoid repeated filtering on subsequent runs
- Download multi-year OHLCV data
- Incrementally update stored datasets
- Compute trend, momentum, volatility indicators
- Store enriched datasets for reuse
- SMA, VWMA
- RSI, Stochastic RSI
- Bollinger Bands
- ATR, TRIX
- MACD (line, signal, histogram)
- ROC, CCI
- DMA and deviation measures
src/
├── universe_selection.py
├── indicator_engineering.py
Gowtham Vuppaladhadiam