AgroFuturesSim is a production-grade financial simulation platform for Argentina’s soybean futures market.
It integrates Hyperledger Fabric blockchain data with advanced stochastic volatility modeling (Heston model) to run large-scale Monte Carlo simulations, produce risk analytics (VaR, CVaR, stress tests), and generate publication-ready reports with Matplotlib-powered charts.
-
🔐 Blockchain Integration (Hyperledger Fabric)
- Securely fetches verified soybean futures trades
- Ensures traceability & auditability with on-chain data validation
-
🧠 Financial Modeling
- Implements Heston stochastic volatility model
- Calibrated against historical / implied volatilities
-
⚡ Simulation Engine
- Monte Carlo (local + distributed with Ray/Dask)
- Checkpointing for long simulations
-
📊 Risk Analytics
- Value at Risk (VaR)
- Conditional VaR (CVaR)
- Extreme stress testing scenarios
-
📑 Automated Reporting
- Multi-page PDF reports (tables, charts, logs)
- Excel summary sheets
- Blockchain provenance metadata for compliance
-
🌐 API + CLI
- FastAPI RESTful endpoints for programmatic workflows
- Click-based CLI for command-line interaction
-
🛡️ Monitoring & Logging
- Centralized logging with rotation
- Prometheus metrics for calibration & simulation progress
┌───────────────────────┐
│ Blockchain Layer │ 🔐 Hyperledger Fabric
└──────────┬────────────┘
│
┌──────────▼────────────┐
│ Data Processing │ ✨ Cleaning, Transform, Cache
└──────────┬────────────┘
│
┌──────────▼────────────┐
│ Modeling Layer │ 📈 Heston Model + Calibration
└──────────┬────────────┘
│
┌──────────▼────────────┐
│ Monte Carlo Engine │ 🎲 Local & Distributed
└──────────┬────────────┘
│
┌──────────▼────────────┐
│ Risk Analytics │ 📊 VaR, CVaR, Stress Testing
└──────────┬────────────┘
│
┌──────────▼────────────┐
│ Visualization/Reports │ 📑 Matplotlib + PDF
└───────────────────────┘
Simulated price paths under calibrated Heston model:
import matplotlib.pyplot as plt
import numpy as np
time = np.linspace(0, 1, 252)
plt.figure(figsize=(10,6))
for i in range(10):
plt.plot(time, np.cumprod(1 + 0.01*np.random.randn(len(time))), lw=1.2)
plt.title("Simulated Soybean Futures Price Paths")
plt.xlabel("Time (Years)")
plt.ylabel("Price")
plt.grid(True)
plt.show()
📌 Example output:
git clone https://github.com/your-org/AgroFuturesSim.git
cd AgroFuturesSim
pip install -r requirements.txt
Fetch blockchain data:
python -m AgroFuturesSim.cli.main fetch-blockchain
Calibrate the Heston model:
python -m AgroFuturesSim.cli.main calibrate-model --config config/default.yaml
Run Monte Carlo simulations:
python -m AgroFuturesSim.cli.main simulate --config config/default.yaml
Generate PDF report:
python -m AgroFuturesSim.cli.main report --output soybean_report.pdf
Start FastAPI server:
uvicorn AgroFuturesSim.api.server:app --reload --host 0.0.0.0 --port 8000
Check system health:
curl http://localhost:8000/health
Calibrate, simulate, and fetch reports via REST endpoints.
Histogram of simulated futures prices:
import matplotlib.pyplot as plt
import numpy as np
plt.figure(figsize=(8,5))
prices = np.random.normal(100, 20, 10000)
plt.hist(prices, bins=50, density=True, alpha=0.7, color="steelblue")
plt.axvline(np.percentile(prices, 5), color="red", linestyle="--", label="VaR (5%)")
plt.title("Monte Carlo Distribution of Soybean Futures")
plt.xlabel("Final Price")
plt.ylabel("Density")
plt.legend()
plt.show()
Build and run:
docker build -t agrofuturessim .
docker run -p 8000:8000 agrofuturessim
Run unit + integration tests:
pytest -v --maxfail=1 --disable-warnings
- Multi-asset Heston model (soybean + corn futures)
- Advanced portfolio optimization under Monte Carlo scenarios
- Real-time ML-based volatility calibration
- Interactive analytics dashboard (Plotly Dash / Streamlit)
Developed by Bhanu Karnwal 🚀
This project is licensed under the Apache License 2.0 - see the LICENSE file for details.