A lightweight, explainable financial risk assessment tool that:
- Simulates future stock price paths using Monte Carlo methods
- Detects volatility regimes via unsupervised clustering
- Estimates short-term drawdown probabilities using rule-based statistical heuristics
Built with real-world market data from Yahoo Finance and visualized in an intuitive Streamlit dashboard.
- 📊 Monte Carlo Simulation: Projects possible stock price paths into the future
- 🌀 Volatility Regime Detection: Labels periods as "stable" or "volatile" using unsupervised clustering
⚠️ Drawdown Risk Estimation: Calculates the likelihood of significant price drops over short horizons- 📈 Interactive Visualization: Streamlit + Plotly interface for exploring simulated risks and market regimes
- Language: Python
- Libraries: NumPy, Pandas, yfinance, scikit-learn, Plotly, Streamlit
- Simulation: Geometric Brownian motion (Monte Carlo)
- Clustering: Regime detection using rolling volatility and thresholds
| File | Description |
|---|---|
app.py |
Main Streamlit dashboard |
monte_carlo_sim.py |
Simulation + volatility clustering module |
tickers.json |
Key-value mapping of stock tickers |
drawdown_risk_classifier.gif |
App demo GIF (used in README) |
requirements.txt |
Python dependencies |
- Clone the repo:
git clone https://github.com/<your-username>/StockVolatility-RiskClassifier.git
cd StockVolatility-RiskClassifier- Install requirements:
pip install -r requirements.txt- Launch the Streamlit app:
streamlit run app.pyWant to estimate the probability of a 15% drawdown in Tesla over the next 30 days? This app simulates 10,000 future paths and returns the short-term risk, contextualized by the current volatility regime.
- Built for practical risk assessment using realistic, interpretable logic
- Ideal for quant research, portfolio monitoring, or market crash detection
- No black-box ML — model behavior is fully explainable and parameterized
This project is licensed under the MIT License.