MSc Economics & Finance, University of Milano-Bicocca — Expected March 2027
I use this GitHub to publish selected coursework and research projects in quantitative finance. The completed projects are mainly in R and MATLAB; Python work will be added as it is developed.
I'm mainly interested in financial time series, risk, derivatives and fixed income, especially in how model assumptions and validation choices affect results.
Group coursework in MATLAB on EURO STOXX 50 options, implied volatility, Monte Carlo pricing, a classroom VSTOXX approximation, EUR OIS bootstrapping and a callable credit-equity-linked bond.
Individual coursework in R across 20 exercises: return diagnostics, ARMA/ARIMA, unit-root testing, ARCH/GARCH/EGARCH, volatility forecasting and VaR/Expected Shortfall. The repository also documents corrections identified during a later review.
Group coursework in R covering Newey-West inference, AR-GARCH simulation, Variance-Gamma marginals with a Gaussian copula, and a CreditRisk+-style loss model with Student-t dependence.
Group coursework in MATLAB comparing equal weighting, risk-based allocation, momentum, trend-following and volatility targeting, with a weekly adaptive-allocation extension.
Group coursework in MATLAB applying the quadratic capital-allocation framework of Dhaene et al. to empirical VaR/TCE and rolling allocation analysis for a five-equity portfolio.
Research design for a thesis on 21-trading-day tail-risk forecasting for SPY. The design compares historical and filtered-historical approaches with a GARCH/EVT specification, then tests whether VIX-based information improves physical VaR/Expected Shortfall forecasts. The forecasts are intended to feed a no-leverage SPY/cash risk-budgeting rule. No empirical results are claimed yet.
Planned Python project on one-month-ahead changes in the Italy-Germany 10-year sovereign spread. The design uses expanding-window evaluation, zero-change and AR(1) benchmarks, Elastic Net and HistGradientBoosting, a market-only versus market-plus-macro comparison, a CISS-based stress slice, and a small DV01 module to translate forecast errors into bond-risk terms. Implementation has not started.
No public exercises yet. I’ll add only self-contained exercises that are useful enough to keep and revisit.
- R — econometrics, simulation, dependence modelling and risk analysis
- MATLAB — derivatives, portfolio allocation and numerical finance
- Python — currently being developed for time-series, ML and fixed-income research
- Bloomberg Terminal — academic market-data work