Development of a real time price forecasting dashboard using Prophet and news sentiment analysis with LLM

Data scientist | Anass MAJJI

---

🧐 Description

This project combines both real-time and historical stock price data with news sentiment analysis to provide more accurate price forecasts and analyze market trends. The sentiment analysis is performed using models like DistilBERT or a quantized version of LLaMA with 1 billion paramters, which analyze stock related news and generate sentiment scores. These sentiment scores are then integrated into one of the price forecasting models such as ARIMA, LSTM, and Prophet.

• By incorporating sentiment analysis as an additional feature, the models can take into account external factors (news sentiment) that may influence stock's future price movements.

• The user can select :

  • News sentiment analysis model: DistilBert or a quantized version of LLaMA with 1B paramters.
  • Price forcasting model such as ARIMA, LSTM or PROPHET.

The data for both historical and real time prices is fetched using the YFinance and NewsApi API.

🚀 Getting Started

1. Clone the repository

git clone https://github.com/your-username/stock-price-forecasting-sentiment-analysis.git
cd stock-price-forecasting-sentiment-analysis

2. Create a virtual environment

python -m venv venv
source venv/bin/activate 

3. Install required dependencies

pip install -r requirements.txt

4. Run the Streamlit Dashboard

streamlit run app.py

🔥 Demo of the Dashboard

Contributing 🤝

Contributions to this project are welcome! Feel free to submit issues or pull requests for improvements.

📪 Contact

For any information, feedback or questions, please [contact me][anass-email]