Gold Price Predictor — Indian Market (MCX)

A full-stack Streamlit application that predicts gold prices (₹ per 10g) using an LSTM deep learning model and provides real-time alert notifications via email.

Dashboard

Alert UI

Received Mail

Features

• 📈 Gold Price Prediction

  • LSTM-based deep learning model
  • Predicts next-day gold prices using historical data

• ⚡ Live Price Tracking

  • Fetches real-time gold price and USD/INR rate
  • Auto-refresh support

• 🔔 Smart Alerts System

  • Price Above / Below alerts
  • Predicted price alerts
  • Email notifications (via Resend API)

• 📊 Interactive Dashboard

  • Actual vs Predicted graph
  • Technical indicators (MA, RSI, Bollinger Bands)
  • Training loss visualization

Tech Stack

• Frontend: Streamlit
• Backend: Python
• ML Model: TensorFlow (LSTM)
• Data Source: Yahoo Finance (yfinance)
• Email Service: Resend API
• Database: Session-based (Streamlit state)

Installation

1. Clone the repository

git clone https://github.com/Asubtlecoderrr/Gold-stock-analysis.git
cd Gold-stock-analysis

2. Create virtual environment

python -m venv venv
source venv/bin/activate   # Mac/Linux
venv\Scripts\activate      # Windows

3. Install dependencies

pip install -r requirements.txt

4. Environment Setup

Create a .env file in the root directory:

RESEND_API_KEY=your_resend_api_key

5. Run the App

streamlit run app.py

Email Alerts (Resend)

• Uses onboarding@resend.dev as default sender
• Only verified emails can receive alerts (in free plan)
• Ensure your email is same as your Resend account

Model Details

• Architecture:

  • LSTM (128 units)
  • Dropout
  • LSTM (64 units)
  • Dense layers

• Input Features:

  • OHLC prices
  • Volume
  • USD/INR
  • Moving averages
  • RSI, Bollinger Bands
  • Volatility

License

This project is for educational purposes only.