How rohannx29/stock-price-prediction-webapp works
This is a Streamlit stock-price and trend prediction dashboard. The principal workflow runs from user-selected ticker and trading settings through Yahoo Finance data acquisition, technical-feature engineering, model training, and an in-range prediction or future forecast; results include charts, evaluation metrics, and downloadable CSVs. Historical queries use as-of retraining, while optional studies compare validation and modeling choices. The graph follows README-described orchestration and shown module calls; some app internals and lookup behavior were not excerpted, so their details are represented only where documented.
Main entry points: app.py Dashboard orchestration · models.py Regression and classification · forecasting.py Recursive forecasts · evaluation.py Metrics and baselines
Components 16
Dashboard
- Dashboard orchestration
app.py - Charts and downloads
app.py
Market data
- OHLCV acquisition
src/data_loader.py - Technical features
src/indicators.py - NYSE calendar
src/calendar_utils.py
Prediction
- Regression and classification
src/models.py - Recursive forecasts
src/forecasting.py - As-of retraining
src/walk_forward.py - Nearest-row lookup
src/lookup.py
Evaluation and studies
- Metrics and baselines
src/evaluation.py - Time-series validation
src/cross_validation.py - Feature ablation
src/feature_ablation.py - Target comparison
src/target_experiment.py - Horizon comparison
src/horizon_study.py
Other components
- Trader
- Yahoo Finance
Connections 27
- Trader → Dashboard orchestration: selects and runs
README.md - Dashboard orchestration → OHLCV acquisition: requests market data
README.md - OHLCV acquisition → Yahoo Finance: downloads OHLCV
src/data_loader.py - Dashboard orchestration → Technical features: builds features
README.md - Dashboard orchestration → Regression and classification: trains and predicts
README.md - Dashboard orchestration → Recursive forecasts: requests future forecast
README.md - Recursive forecasts → Technical features: recomputes features
src/forecasting.py - Recursive forecasts → Regression and classification: predicts next step
src/forecasting.py - Recursive forecasts → NYSE calendar: advances trading day
src/forecasting.py - Dashboard orchestration → As-of retraining: retrains for history
README.md - As-of retraining → Regression and classification: trains as-of models
src/walk_forward.py - Dashboard orchestration → Nearest-row lookup: locates requested row
README.md - Dashboard orchestration → Metrics and baselines: reports metrics
README.md - Dashboard orchestration → Time-series validation: runs optional validation
app.py - Time-series validation → Regression and classification: trains and predicts
src/cross_validation.py - Time-series validation → Metrics and baselines: scores folds
src/cross_validation.py - Dashboard orchestration → Feature ablation: runs optional study
app.py - Feature ablation → Regression and classification: trains feature variants
src/feature_ablation.py - Feature ablation → Metrics and baselines: scores variants
src/feature_ablation.py - Dashboard orchestration → Target comparison: runs optional comparison
app.py - Target comparison → Regression and classification: trains target variants
src/target_experiment.py - Target comparison → Metrics and baselines: scores targets
src/target_experiment.py - Dashboard orchestration → Horizon comparison: runs optional comparison
app.py - Horizon comparison → Recursive forecasts: evaluates recursive strategy
src/horizon_study.py - Horizon comparison → Regression and classification: trains direct models
src/horizon_study.py - Horizon comparison → Metrics and baselines: scores horizons
src/horizon_study.py - Dashboard orchestration → Charts and downloads: renders and exports
README.md