Financial Market Intelligence Platform

Overview

A tool for analyzing financial markets using deep learning. This platform predicts stock price movements with an LSTM model and analyzes market sentiment from news articles using FinBERT. It's all wrapped up in an interactive Streamlit dashboard and a FastAPI backend.

Features

• Stock Price Prediction: LSTM model to forecast future stock prices.
• Sentiment Analysis: FinBERT model to analyze financial news sentiment.
• Interactive Dashboard: Streamlit app for visualizing data and predictions.
• REST API: FastAPI backend to serve model predictions.
• Containerized: Docker support for easy setup.

Tech Stack

Category	Technology
Backend	FastAPI, Uvicorn
Frontend	Streamlit
Deep Learning	PyTorch, Hugging Face Transformers
Data	Pandas, NumPy, Scikit-learn, yfinance
Deployment	Docker

Getting Started

Installation

1. Clone the repo

   git clone https://github.com/LUCKYREDDY31/financial-market-intelligence.git
   cd financial-market-intelligence

2. Run the setup script

   ./setup.sh

   This will create a virtual environment, install dependencies, download data, and train the models.

3. Activate the environment

   source venv/bin/activate

Usage

• Run the dashboard

  streamlit run app/dashboard.py

  The app will be available at http://localhost:8501.

• Run the API

  python src/api/main.py

  The API will be available at http://localhost:8000 with docs at /docs.

Docker

1. Build the image

   docker build -t market-intelligence .

2. Run the container

   docker run -p 8501:8501 market-intelligence

   The app will be available at http://localhost:8501.

Project Structure

financial-market-intelligence/
├── app/                    # Streamlit dashboard
├── src/
│   ├── api/               # FastAPI backend
│   ├── data_processing/   # Data download & features
│   ├── models/            # LSTM & FinBERT
│   └── train_models.py    # Training script
├── data/                  # Data storage
└── Dockerfile             # For deployment

How It Works

1. Data Ingestion: Historical stock data is downloaded from Yahoo Finance using yfinance.
2. Feature Engineering: Technical indicators (RSI, MACD, etc.) are calculated to create features for the LSTM model.
3. Model Training: The LSTM model is trained on the historical data to predict future prices. The FinBERT model is used for sentiment analysis on news headlines.
4. API & Dashboard: The trained models are served via a FastAPI backend and visualized in a Streamlit dashboard.

Disclaimer

This project is for educational purposes only and is not financial advice. Do not use it for making real investment decisions.

Contributing

Contributions are welcome! Please open an issue or submit a pull request.

License

This project is licensed under the MIT License.