🚀 Algo-Trading System with ML, Automation, and Live Alerts

An end-to-end algorithmic trading prototype built with Python, powered by RSI + DMA strategy, enhanced with ML predictions, Google Sheets P&L tracking, Groq commentary, and Telegram alerts — all wrapped in a stylish Flask UI.

---

🔧 Tech Stack

Component	Stack / Tool
Strategy Logic	RSI + 20DMA / 50DMA Crossover
ML Model	Scikit-Learn (Decision Tree)
Stock Data	Yahoo Finance via yfinance
Automation	Python + Flask
Logging & P&L	Google Sheets API + GCP Service
Alerts	Telegram Bot API
Commentary	Groq API (LLaMA3-70B Instruct)
UI	HTML5 + CSS3 (Glassmorphism style)

---

📈 Strategy Logic

This project implements a hybrid technical + rule-based strategy:

• ✅ Buy Signal Conditions:

  • RSI < 30 (oversold)
  • 20DMA crosses above 50DMA
  • Optional relaxed condition: RSI < 50 + 20DMA > 50DMA (for testing)

• ✅ Backtest Window:

  • Last 6 months (period='6mo') using daily candles (interval='1d')

• ✅ Backtest Output:

  • Total Trades
  • Win Count / Loss Count
  • Win Ratio
  • Total P&L

---

🤖 Machine Learning Automation

• Model: DecisionTreeClassifier
• Features Used:
  • RSI
  • 20DMA
  • 50DMA
  • MACD
  • Volume
• Target: Whether next-day close is higher than today’s close
• Train/Test Split: 80/20
• Accuracy Output is displayed on the UI

---

☁️ Google Sheets Logging (via GCP)

Trade and summary logs are stored live using:

• 🔐 A creds.json file (GCP Service Account key)
• 📊 Trade logs go to Trade_Log sheet
• 📘 Summary stats go to Summary tab with:
  • Date
  • Total Trades
  • Wins / Losses
  • Win Ratio
  • P&L

Libraries Used:

• gspread
• oauth2client

---

📬 Telegram Alerts

Each trade signal is pushed via Telegram Bot API in real-time:

• Uses: requests.post to the Telegram API
• Requires:
  • TELEGRAM_BOT_TOKEN
  • TELEGRAM_CHAT_ID

You’ll get alerts like:
📢 RELIANCE.NS BUY at 2953.25

---

🧠 Groq AI Commentary

Each trading cycle ends with Groq LLaMA3-based summarization:

• Model: llama3-70b-8192
• API: https://api.groq.com/openai/v1/chat/completions
• Output: Smart summary of the day’s trade signals

Example:

“Today, RELIANCE and TCS showed promising RSI-based buy signals supported by DMA crossovers. INFY remained neutral.”

---

💻 Flask UI (Modern 3D-Styled)

✅ Features:

• Responsive, animated UI
• Glassmorphism cards
• Color-coded trade signals
• Display of:
  • Trade Signal Table
  • ML Model Accuracy
  • Groq Commentary

📸 Preview


UI styled using custom CSS with:

• backdrop-filter + -webkit-backdrop-filter for Safari support
• Neon headings
• Card animations and hover effects

---

🗂️ Project Structure

│ ├── app.py # Flask server ├── strategy.py # RSI + DMA logic ├── ml_model.py # Decision Tree model ├── data_fetcher.py # Yahoo Finance fetcher ├── telegram_alert.py # Sends alerts ├── groq_commentary.py # Groq LLM summary ├── google_sheets.py # Google Sheets logging via GCP

│ ├── static/ │ └── style.css # UI styling (glassmorphism) │

├── templates/ │ └── index.html # Frontend dashboard │

├── creds.json # GCP service account key └── .env # Secrets (API keys, tokens)

---

🔑 Environment Setup

Create a .env file with:

GROQ_API_KEY=your_groq_api_key
GOOGLE_SHEET_NAME=YourSheetName
TELEGRAM_BOT_TOKEN=your_bot_token
TELEGRAM_CHAT_ID=your_chat_id

🚀 How to Run

# Step 1: Install dependencies
pip install -r requirements.txt

# Step 2: Run the Flask app
python app.py

# Visit the dashboard at:
http://localhost:5000

---

✅ 10. Demo Videos Section

## 📹 Demo Videos

| Demo Type                | Link                         |
|--------------------------|------------------------------|
| 📽️ Strategy & Code Flow | [Watch on Google Drive](#)   |
| 📊 Output + Google Sheets | [Watch on Google Drive](#)  |

✅ Deliverables Checklist

• ✅ RSI + DMA strategy logic
• ✅ 6-month backtesting
• ✅ Decision Tree prediction
• ✅ Google Sheets automation
• ✅ Telegram alerts
• ✅ Groq-generated insights
• ✅ Modern web UI
• ✅ Modular, well-documented code

🙌 Author

Shrutakeerti — For any updates please contract through the given links 🔗 GitHub    📬 Telegram    💼 LinkedIn

📃 License

MIT License – use this as a template, study tool, or launchpad for your own bots.

“Systems trade logic. Traders trade emotion. Bots don’t care.”