Python Stock Prediction Model Suite

I need a complete, reproducible workflow in Python that forecasts daily stock closing prices and short-term direction. Historical data should be pulled directly from Yahoo Finance, then enriched with the key technical indicators that matter most to me—moving averages, RSI, and MACD—along with any other sensible features you believe will help the model generalise without leaking future information.

The core of the project is a time-series experiment: build and compare both LSTM and RNN architectures against simpler baselines such as Linear Regression, Random Forest, and XGBoost. Please handle the train/validation/test split with proper chronological order so no look-ahead bias sneaks in. I’ll judge performance with RMSE, MAE, and directional accuracy.

To make results easy to digest, include clear Matplotlib/Seaborn visualisations: actual-versus-predicted price curves, indicator overlays, and loss trajectories for each network. I prefer the work delivered as a well-commented Jupyter Notebook, but separate .py modules for reusable components are fine.

Deliverables
• Clean, documented Python code (Notebook + modules)
• Feature-engineering functions for moving averages, RSI, MACD, etc.
• Trained LSTM, RNN, and baseline model objects with saved weights or joblib files
• Plots and metric tables demonstrating model performance
• Brief write-up (Markdown or PDF) describing data prep, modelling choices, and results
• (Bonus) Lightweight Streamlit or Flask dashboard that pulls the latest prices and shows live predictions

I’m happy to clarify objectives or data nuances as you go—efficient, well-explained work is what I value most.