Tesla Stock Price Prediction (Zero to Deployment)

Overview

This project builds a time series forecasting model to predict Tesla's stock prices using LSTM (Long Short-Term Memory) networks. The model is trained on historical stock data and deployed as an API using FastAPI with Uvicorn.

Features

• Data Collection: Fetching historical Tesla stock prices.
• Data Preprocessing: Normalization, feature engineering, and sequence generation.
• Model Training: Implementing LSTM with TensorFlow/Keras.
• Evaluation & Optimization: Hyperparameter tuning and performance analysis.
• Deployment: Serving predictions using FastAPI.

Dataset

• Source: Yahoo Finance
• Data: Tesla (TSLA) stock prices
• Features: Date, Open, High, Low, Close, Volume

Tech Stack

• Python
• TensorFlow & Keras
• Pandas & NumPy
• Scikit-learn
• Matplotlib & Seaborn (for visualization)
• FastAPI & Uvicorn (for API deployment)
• Docker (optional for containerization)

Installation

1. Clone this repository:

   git clone https://github.com/yourusername/Stock_Prediction.git
   cd Stock_Prediction

2. Install dependencies:

   pip install -r requirements.txt

Model Training

Run the training script:

python train.py

This will:

• Load and preprocess the Tesla stock dataset
• Train an LSTM model
• Save the trained model as model.h5

Deployment

1. Start the FastAPI server:

   uvicorn app:app --reload

2. API Endpoint:

   GET /predict?days=5

   Example request:

   curl -X GET "http://127.0.0.1:8000/predict?days=5"

   Example response:

   {
     "predicted_prices": [880.45, 890.12, 905.67, 915.32, 925.89]
   }

Future Improvements

• Integrate Sentiment Analysis on news headlines for better forecasting.
• Implement GRU or Transformer-based models for comparison.
• Deploy on Cloud (AWS/GCP/Azure) for real-time inference.

Contributing

Feel free to fork this repository and submit a pull request if you have any improvements!

License

This project is licensed under the MIT License.