How rajanguptashah/stock_price_forecasting works
Project type and purpose
• This is a data-driven, full-stack forecasting application written in Python.
• Purpose: ingest historical stock data (via Stooq API or offline CSV), engineer features (RSI, EMA, MACD), train a time-series forecasting model (Prophet), evaluate performance, and expose results through both a CLI and an interactive Dash web dashboard.
File structure
• Top level:
• Root CSV files (AAPL.csv, AMZN.csv, etc.) for offline mode
• stock_forecast.py – main CLI script that orchestrates data fetch, feature engineering, modeling, evaluation, and static plots
• (Implicit) app.py – Dash server for the interactive dashboard
• tempCodeRunnerFile.py – scratch or auxiliary code
• No separate “src/” or “frontend/” folders—this is a monolith that bundles data ingestion, ML pipeline, and dashboard.
• Configuration and dependency manifest implied via requirements.txt
README for additional insights
• Data acquisition: Stooq API with offline fallback to CSV.
• Technical indicators: RSI, EMA-9, EMA-21, MACD.
• Forecasting engine: Facebook Prophet.
• Visualization: Plotly (static) and Dash (interactive).
• Evaluation metrics: MAE, RMSE, MAPE.
Final advice
• Conditional logic for offline/online data fetching.
• Both CLI and web UIs so that stakeholders understand dual access modes.