🚀 Crypto Trading Bot

A Python-based crypto trading bot that combines technical analysis, machine learning, grid trading, and sentiment analysis to make automated trading decisions.

---

📌 Overview

This bot monitors a configured trading pair, evaluates market conditions using multiple signals, and executes trades when all conditions align.

It integrates:

• 📊 Technical Indicators (SMA, EMA, RSI, MACD, Bollinger Bands)
• 🤖 Machine Learning (RandomForestClassifier)
• 📉 Grid Trading Strategy
• 🐦 Twitter/X Sentiment Analysis

---

✨ Features

• Fetches OHLCV market data using ccxt
• Computes key technical indicators using ta
• Trains a machine learning model on recent market data
• Implements a grid-based trading strategy
• Uses Twitter sentiment as an additional signal filter
• Logs all activity for monitoring and debugging
• Stores trade state locally

---

🧠 How It Works

1. Loads API credentials from .env

2. Loads trading configuration from config.json

3. Connects to the exchange

4. Fetches market data and computes indicators

5. Trains ML model on historical data

6. Generates grid levels

7. Runs continuous loop:

   • Fetch latest data
   • Recalculate indicators
   • Predict market movement
   • Analyze sentiment
   • Execute buy/sell orders if conditions match

---

🏗️ Project Structure

bot.py                 # Main trading loop
data_fetch.py          # Exchange connection & data retrieval
indicators.py          # Technical indicators
grid_strategy.py       # Grid trading logic
ml_strategy.py         # ML model training & prediction
twitter_sentiment.py   # Sentiment analysis
state_manager.py       # Trade state handling
utils.py               # Helper functions
config.json            # Bot configuration
requirements.txt       # Dependencies

---

⚙️ Requirements

• Python 3.10+
• Exchange API credentials (e.g., Binance)
• Twitter/X Bearer Token

Install dependencies:

pip install -r requirements.txt

---

🔐 Configuration

1. Environment Variables

Create a .env file:

BINANCE_API_KEY=your_api_key
BINANCE_API_SECRET=your_api_secret
TWITTER_BEARER_TOKEN=your_twitter_bearer_token

⚠️ Important:

• Never commit real credentials
• Use minimal permissions
• Prefer testnet/paper trading first

---

2. Bot Settings (config.json)

{
  "exchange": "binance",
  "symbol": "BTC/USDT",
  "timeframe": "1h",
  "grid_levels": 10,
  "grid_step_pct": 0.5,
  "order_size": 0.001,
  "base_currency": "USDT",
  "twitter": {
    "query": "BTC OR Bitcoin -is:retweet lang:en",
    "max_tweets": 80
  }
}

---

▶️ Running the Bot

python bot.py

---

📊 Strategy Logic

The bot uses three filters together:

1. Grid Condition

Trades only when price is near predefined grid levels.

2. ML Condition

• Buy → High probability of upward movement
• Sell → Low probability

3. Sentiment Condition

• Buy → Positive sentiment
• Sell → Negative sentiment

Default thresholds:

Buy  → prob_up > 0.55 AND sentiment > 0.1
Sell → prob_up < 0.45 AND sentiment < -0.1

---

📁 Outputs

• logs/bot.log → Logs, signals, and errors
• trade_state.json → Trade history/state
• ml_model.pkl → Saved ML model

---

⚠️ Limitations

• Executes real market orders
• No backtesting pipeline yet
• ML model is basic (no validation)
• Limited error handling (fees, slippage, API failures)
• Sentiment model may be slow on first run

---

🔮 Future Improvements

• Add backtesting module
• Support testnet/paper trading
• Improve risk management
• Save scaler/model pipeline
• Add alerts & monitoring
• Write unit tests

---

📌 Disclaimer

This project is for educational purposes only. Cryptocurrency trading is highly risky. Use at your own risk.

---

💡 Tech Stack

• Python
• ccxt
• ta (Technical Analysis)
• scikit-learn
• Transformers (for sentiment analysis)

---

---