🔮📈 Predict‑Crypto

A machine‑learning project to analyze and forecast cryptocurrency prices (Bitcoin & Dogecoin) using historical market data, enriched feature sets, and time‑series models (ARIMA/SARIMA).

---

📋 Table of Contents

• 🔮📈 Predict‑Crypto
  • 📋 Table of Contents
  • 🚀 Features
  • 📂 Repository Structure
  • 📈 Model Results
    • Dogecoin: Regression Performance by Forecast Horizon
  • 🗂️ Project Status
  • 🔮 Future Plans
  • ⭐ Contributing & Feedback

---

🚀 Features

• 📊 Time‑series forecasting: ARIMA/SARIMA baseline models
• 🔍 Feature engineering: Lag features, rolling statistics, VWAP
• 🛠️ Extensible code: Shared utilities in src/
• 📒 Interactive notebooks for Bitcoin & Dogecoin
• ☁️ BigQuery integration for raw blockchain data

---

📂 Repository Structure

📦 root/
├── 📁 data/
│   ├── 📁 bitcoin/
│   └── 📁 doge/
├── 📁 notebooks/
│   ├── 📁 bitcoin/
│   └── 📁 doge/
├── 📁 src/
├── 📁 img/
├── 📁 bigquery/
└── 📄 README.md

---

📈 Model Results

Dogecoin: Regression Performance by Forecast Horizon

Model	1‑Day MSE	1‑Day R²	2‑Day MSE	2‑Day R²	3‑Day MSE	3‑Day R²	4‑Day MSE	4‑Day R²
Linear Regression	0.000101	0.9842	0.000186	0.9710	0.000273	0.9574	0.000372	0.9421
Random Forest	0.000724	0.8873	0.001724	0.7313	0.003403	0.4703	0.005207	0.1894
XGBoost	0.001445	0.7749	0.001115	0.8262	0.001804	0.7192	0.003077	0.5210
LightGBM	0.000504	0.9215	0.000717	0.8884	0.001754	0.7269	0.003881	0.3959

Deep dive & narrative:
Read the Medium article When Simplicity Wins….

📓 Notebook: notebooks/doge/predict_OHLCVM.ipynb

---

🗂️ Project Status

Component	Status	Notes
Data Collection	🔄 In Progress	Doge: 2013–2025 daily; BTC: minute‑level
Feature Engineering	🔄 In Progress	1‑day lags, 7‑day rolls, VWAP
ARIMA/SARIMA Baselines	🔄 In Progress	Hyperparameter search in progress

---

🔮 Future Plans

1. Model Expansion
   • ARIMA/SARIMA, Prophet, LSTM/TCN/Transformer-based forecasters
2. Advanced Features
   • Technical indicators (RSI, MACD), sentiment analysis
3. Robust Validation
   • Walk‑forward CV, backtesting on event‑driven windows
4. Deployment
   • Streamlit/Dash app for live updates; GitHub Actions for retraining
5. Data Enrichment
   • On‑chain metrics, order‑book snapshots; scalable DB backend

---

⭐ Contributing & Feedback

Feel free to:

• ⭐ Star this repo if you find it useful
• 🐛 Report issues or 📥 submit PRs
• 💬 Reach out on LinkedIn for collaboration

---

Last updated: July 2025