jo-marcel/spy-volatility-research

How jo-marcel/spy-volatility-research works

This is a quantitative research workflow studying whether trailing SPY volatility predicts next-day absolute-return magnitude—not a trading system. The principal run loads prices, builds features, analyzes full-sample association, evaluates fixed holdout and yearly walk-forward predictions, analyzes volatility regimes, then produces charts and a summary. The graph preserves documented model, baseline, metrics, visualization, and regime stages. Edges follow the main and sampled module calls; some documented capabilities have unsampled implementations, so their internal links are omitted.

Main entry points: main.py Study Runner · holdout.py Fixed Holdout · walk_forward.py Walk-Forward Evaluation · predict.py Model Predictions

Components 20

Data and Features

Shared Settings config.py
SPY Prices SPY_2015_2025.csv
Price Loader data_loader.py
Feature Builder features.py

Analysis and Evaluation

Study Runner main.py
Association Analysis inference.py
Fixed Holdout holdout.py
Chronological Splits split_data.py
Model Predictions predict.py
Regression Models models.py
Mean Median Baselines baselines.py
Error Metrics metrics.py
Walk-Forward Evaluation walk_forward.py
Volatility Regimes regimes.py

Research Outputs

Historical Charts plots_history.py
Model Charts plots_models.py
Summary Export export_summary.py
Research Figures figures
Research Results results/summary.json

Other components

Researcher

Connections 28

Researcher → Study Runner: runs README.md
Study Runner → Price Loader: loads prices main.py
Price Loader → SPY Prices: reads data_loader.py
Price Loader → Shared Settings: uses path data_loader.py
Study Runner → Feature Builder: builds features main.py
Feature Builder → Shared Settings: uses settings features.py
Study Runner → Association Analysis: analyzes association main.py
Association Analysis → Regression Models: uses design matrix inference.py
Study Runner → Fixed Holdout: runs holdout main.py
Fixed Holdout → Chronological Splits: splits data holdout.py
Fixed Holdout → Model Predictions: predicts holdout.py
Fixed Holdout → Error Metrics: scores holdout.py
Fixed Holdout → Research Results: writes predictions and metrics holdout.py
Model Predictions → Regression Models: fits and predicts predict.py
Model Predictions → Mean Median Baselines: adds baselines predict.py
Study Runner → Walk-Forward Evaluation: runs yearly evaluation main.py
Walk-Forward Evaluation → Chronological Splits: splits by year walk_forward.py
Walk-Forward Evaluation → Model Predictions: predicts yearly walk_forward.py
Walk-Forward Evaluation → Error Metrics: scores yearly walk_forward.py
Walk-Forward Evaluation → Research Results: writes predictions and metrics walk_forward.py
Study Runner → Volatility Regimes: analyzes regimes main.py
Study Runner → Historical Charts: plots history main.py
Study Runner → Model Charts: plots model results main.py
Historical Charts → Shared Settings: uses figure path plots_history.py
Historical Charts → Research Figures: writes charts plots_history.py
Study Runner → Summary Export: exports summary main.py
Summary Export → Shared Settings: uses results path export_summary.py
Summary Export → Research Results: writes summary export_summary.py

Drawn by GitDiagram on November 12, 2026 from the repository's file tree, README and source files. Read as Markdown ·

How jo-marcel/spy-volatility-research works

Main entry points: main.py Study Runner · holdout.py Fixed Holdout · walk_forward.py Walk-Forward Evaluation · predict.py Model Predictions

Connections 28

Researcher → Study Runner: runs README.md
Study Runner → Price Loader: loads prices main.py
Price Loader → SPY Prices: reads data_loader.py
Price Loader → Shared Settings: uses path data_loader.py
Study Runner → Feature Builder: builds features main.py
Feature Builder → Shared Settings: uses settings features.py
Study Runner → Association Analysis: analyzes association main.py
Association Analysis → Regression Models: uses design matrix inference.py
Study Runner → Fixed Holdout: runs holdout main.py
Fixed Holdout → Chronological Splits: splits data holdout.py
Fixed Holdout → Model Predictions: predicts holdout.py
Fixed Holdout → Error Metrics: scores holdout.py
Fixed Holdout → Research Results: writes predictions and metrics holdout.py
Model Predictions → Regression Models: fits and predicts predict.py
Model Predictions → Mean Median Baselines: adds baselines predict.py
Study Runner → Walk-Forward Evaluation: runs yearly evaluation main.py
Walk-Forward Evaluation → Chronological Splits: splits by year walk_forward.py
Walk-Forward Evaluation → Model Predictions: predicts yearly walk_forward.py
Walk-Forward Evaluation → Error Metrics: scores yearly walk_forward.py
Walk-Forward Evaluation → Research Results: writes predictions and metrics walk_forward.py
Study Runner → Volatility Regimes: analyzes regimes main.py
Study Runner → Historical Charts: plots history main.py
Study Runner → Model Charts: plots model results main.py
Historical Charts → Shared Settings: uses figure path plots_history.py
Historical Charts → Research Figures: writes charts plots_history.py
Study Runner → Summary Export: exports summary main.py
Summary Export → Shared Settings: uses results path export_summary.py
Summary Export → Research Results: writes summary export_summary.py

Drawn by GitDiagram on November 12, 2026 from the repository's file tree, README and source files. Read as Markdown ·