Quantitative Risk Management in Python
Learn to measure, simulate and hedge portfolio risk in Python with VaR, CVaR, Monte Carlo simulation and extreme value theory, using the 2007-2008 crisis.
Quantitative risk management in Python is taught here through the 2007-2008 financial crisis and what it did to investment banks such as Goldman Sachs and J.P. Morgan. You measure a portfolio’s exposure with Value at Risk and Conditional Value at Risk, estimate losses with parametric and Monte Carlo methods, and learn to spot structural breaks that show when market randomness itself changed. At its center is a problem shared by banking, insurance and asset management: balancing reward against exposure to risk.
What You Will Achieve
- Calculate Value at Risk and Conditional Value at Risk for a portfolio and compare what each says about potential loss
- Offset the risk of an options portfolio using prices from the Black-Scholes model
- Simulate possible portfolio losses with Monte Carlo methods and real historical data
- Model extreme losses with GEV and kernel density estimates, then choose loss cover with CVaR
Program Overview
Risk, Return and Modern Portfolio Theory
A recap of how risk and return relate, which risk factors drove the crisis, and how portfolio theory reads that period.
- Portfolio returns and volatility during the crisis
- Risk factors and a least-squares factor model
- PyPortfolioOpt practice and the efficient frontier
Risk Measures and Options Hedging
Goal-oriented tools that tie risk measures to exposure and appetite, then use options to offset part of the risk.
- VaR and CVaR under normal and fitted distributions
- Risk exposure, loss and risk appetite
- Minimizing CVaR in a crisis portfolio
- Black-Scholes pricing and options hedging
Estimating Risk, Extreme Events and Neural Networks
Estimation methods, signs that the crisis changed market behavior, and more general tools for extreme events and complex loss distributions.
- Parametric, historical and Monte Carlo estimation
- Structural breaks, volatility and backtesting
- Extreme value theory and kernel density estimation
- Neural networks for loss distributions and real-time portfolio optimization
Certificate Awarded (Paid)
Job Outlook
These methods apply to portfolio risk work in banking, insurance and asset management, where financial risk analysts, regulators and actuaries measure exposure to loss and weigh it against expected rewards.
Good for: Analysts who want to measure, simulate and hedge portfolio risk in Python, with one historical crisis serving as the running case throughout.
Keep in mind: Every chapter returns to the 2007-2008 financial crisis as its running case, and the neural network lessons use single-layer networks.
Bottom line: A crisis-driven walk through portfolio risk in Python that covers VaR, CVaR, Monte Carlo simulation, structural breaks and extreme value theory.
Frequently Asked Questions
Which Python libraries do the exercises rely on?
The risk measure chapter uses pandas, scipy and pypfopt to compute VaR and CVaR and to minimize CVaR. The opening chapter already practices with PyPortfolioOpt on returns and covariance before turning to the efficient frontier.
How is the 2007-2008 financial crisis used in the lessons?
The crisis is the running case. Lessons look at portfolio returns and risk factors during that period, manage CVaR against it, test for structural breaks around it and study its extreme losses with block maxima and GEV estimation.
How does the course handle structural breaks?
Structural breaks are presented as the sign that randomness itself was changing during the crisis. A series of exercises identifies the break in crisis data, then links it to volatility, extreme values and backtesting.
What are neural networks used for in this course?
The final chapter builds single-layer neural networks to approximate loss distributions, predict asset prices and carry out real-time portfolio optimization, presented as a more general risk management tool alongside extreme value theory and kernel density estimation.


There are no reviews yet.