Monte Carlo simulation surfaced this weight mix as the efficient frontier tangent portfolio, delivering the best risk-adjusted return.
- Return
- 8.2%
- Risk
- 10.2%
- Sharpe
- 0.34
- Beta
- 0.92
- Jensen’s α
- 1.4%
Project Case StudyPortfolio Optimization
Portfolio Optimization Lab
Phase II of my graduate research dives into risk-adjusted portfolio construction. Using R (tidyquant, PortfolioAnalytics, PerformanceAnalytics), I simulated 5,000 Monte Carlo portfolios, analyzed factor exposure, and benchmarked custom strategies against the S&P 500. This dashboard distills the findings into an interactive investor-ready deliverable.
Skills: R · Monte Carlo Simulation · Factor Analytics · Data VisualizationYear: 2025
← Back to all projectsResearch focus
Modern portfolio theory brought to life
Recruiters often ask how I approach portfolio construction. This project answers with data: efficient frontier simulations, Sharpe improvements, S&P outperformance months, and factor regression insights that isolate true alpha.
Portfolio AnalyticsNew Highlight
Portfolio Optimization Lab
This project distills the insights from my “Financial Modeling” coursework and @Phase2_Group1 research. Using R (tidyquant, PortfolioAnalytics, PerformanceAnalytics), I simulated 5,000 allocations across 24 stocks, mapped the efficient frontier, and benchmarked optimized, naive, and thematic portfolios versus the S&P 500.
Efficient Frontier & Scenario Comparison
Each point = simulated portfolioEfficient Frontier (Risk vs Return)
Benchmark portfolio distributing capital evenly across the 24 names to contrast disciplined optimization with naive allocation.
- Return
- 7.3%
- Risk
- 11.8%
- Sharpe
- 0.27
- Beta
- 0.99
- Jensen’s α
- 0.4%
Hand-crafted weights reflect sector tilts and conviction bets; backtests show consistent alpha after controlling for factor exposures.
- Return
- 8.7%
- Risk
- 12.5%
- Sharpe
- 0.31
- Beta
- 1.08
- Jensen’s α
- 1.1%
Used as the market proxy to measure excess performance, beta, and Jensen’s alpha across optimized and heuristic allocations.
- Return
- 6.9%
- Risk
- 10.9%
- Sharpe
- 0.25
- Beta
- 1.00
- Jensen’s α
- 0.0%
Factor Exposures (Fama-French 5)
Regression coefficients vs. S&P 500| Factor | Optimal | Naive | Custom |
|---|---|---|---|
| Market (MKT) | 0.92 | 1.01 | 1.08 |
| Size (SMB) | -0.08 | 0.04 | 0.11 |
| Value (HML) | -0.05 | 0.07 | 0.02 |
| Profitability (RMW) | 0.06 | 0.01 | 0.12 |
| Investment (CMA) | -0.03 | -0.02 | 0.05 |
Notebook Highlights
- 24-security universe spanning tech, consumer staples, industrials, and financials.
- 5,000 Monte Carlo portfolios evaluated monthly using Yahoo Finance data from 2010–2025.
- Factor regressions (CAPM + Fama-French 5) reveal how tilts drive alpha beyond market beta.
- Interactive dashboards compare optimized, naive, and thematic strategies against the S&P 500.
Implementation Notes
- Translated R outputs (`Phase2_Group1.R`) into a reusable TypeScript model (`lib/portfolioLab.ts`) capturing the efficient frontier, highlighted portfolios, and factor loadings.
- Designed the `PortfolioLab` component with Recharts to visualise frontier trade-offs, KPI cards, and factor tables in a finance-friendly aesthetic.
- Focused on narrative clarity: recruiters can instantly see how optimized, naive, and custom allocations differ in risk, return, and alpha.