Quant-grade intelligence for decoding hidden risk premia in crypto and is fully reproducible, noise-engineered, and built for serious researchers and traders seeking alpha beyond the hype.
Crypto markets are wild, noisy, and poorly mapped — a frontier where meaningful factor signals are buried under extreme volatility and speculative noise. Traditional equity factor models don’t translate cleanly, leaving a gap between theory and actionable insight.
This dashboard closes that gap by delivering:
- Clean, factor-based analytics that extract systematic drivers of crypto returns.
- Noise-aware filtering to reveal true premia signals masked by daily chaos.
- Reproducible, research-ready frameworks ideal for academic studies, quant funds, and high-level strategy design.
- A single interactive environment that unifies market, momentum, low-volatility, and network-value factors — and lets you extend to new ones as the space evolves.
- Fetches and cleans OHLCV data for major cryptos (BTC, ETH, altcoins) via yfinance
- Applies rolling median, z-score clipping, EMA smoothing, and more to reduce noise
- Computes market, momentum, low-volatility, network value, and custom factors
- Multi-factor backtest integration (momentum + low-volatility)
- Interactive Streamlit dashboard with raw vs denoised data comparisons
- Expanded dashboard pages: Market Risk Premium, Momentum, Low Volatility, Network Value, Factor Portfolio
- Dark institutional theme for quant feel
- Extensible: add new factors, filters, or data sources easily
- Source: Yahoo Finance (yfinance) for daily/hourly OHLCV
- Noise Reduction: Rolling median, z-score clipping, EMA, optional Kalman filter
- Factor Computation: Market premium, momentum, low-volatility anomaly, etc.
- Visualization: Streamlit dashboard with toggles for raw/cleaned data
- Clone the repo
- Install dependencies:
pip install -r requirements.txt - Run the dashboard:
streamlit run dashboard.py
data_loader.py: Data fetching and cleaningfactors/: Factor computation modulesdashboard.py: Streamlit dashboard
- Sydney Quantitative Finance Symposium, 2023: 'Noise Reduction in Crypto Factor Models'
- EPFL Blockchain Analytics, 2025: 'Analyzing the Predictability of Crypto Markets'
🔜 Add more advanced noise reduction (Kalman, wavelets)
🔜 Factor correlation heatmaps & regime detection
🔜 Machine learning–driven factor forecasts
🔜 Integration with DeFi metrics (on-chain activity, TVL factors)
🔜 Portfolio optimizer with transaction cost modeling
For quant students, researchers, and funds seeking robust, noise-aware crypto analytics.
This project is intended solely for educational purposes and as an innovative guide for quantitative researchers. It does not constitute investment advice or a recommendation to buy, sell, or hold any financial asset. Users should conduct their own due diligence and consult professional advisors before making investment decisions.