Advanced Portfolio Construction with Python Course

Lesson 1 • Computing Returns and Basic Statistics

Calculate simple and log returns and summarise their distributions. These metrics are the direct inputs to all risk and optimisation models ahead.

Lesson 2 • Fetching and Cleaning Market Data

Retrieve historical price data via APIs and clean it for analysis. Ensures data quality before any return or risk calculation is performed.

Lesson 3 • Pandas for Time Series Data

Manipulate DataFrames and time-indexed Series to handle price and return data. Provides the data layer that feeds every subsequent analytical module.

Lesson 4 • Setting Up the Quant Environment

Configure Python, virtual environments, and essential libraries for portfolio analysis. Establishes the reproducible workspace used throughout the course.

Lesson 1 • Covariance Matrix Estimation

Estimate and shrink covariance matrices to reduce estimation error. Reliable covariance matrices are essential for stable portfolio optimisation.

Lesson 2 • Volatility Estimation Techniques

Estimate historical and exponentially weighted volatility from return series. Accurate volatility estimates underpin every risk-adjusted performance measure.

Lesson 3 • Drawdown and Tail Risk Metrics

Measure peak-to-trough losses and extreme downside risk using Python. Connects volatility analysis to investor-relevant loss scenarios.

Lesson 4 • Risk Attribution and Reporting

Attribute portfolio variance to individual assets and factors. Translates raw risk numbers into actionable insights for portfolio managers.

Lesson 1 • Markowitz Framework Essentials

Formalise the mean-variance objective and constraint structure mathematically. Grounds all subsequent optimisation code in rigorous theory.

Lesson 2 • Incorporating Practical Constraints

Add long-only, sector, and turnover constraints to the optimizer. Bridges the gap between textbook theory and real-world portfolio mandates.

Lesson 3 • Resampling and Robust Optimisation

Apply Michaud resampling and robust methods to stabilise optimal weights. Addresses the instability of classical MVO under estimation error.

Lesson 4 • Mapping the Efficient Frontier

Sweep target return levels to trace the full efficient frontier curve. Visualises the risk-return trade-off central to portfolio construction decisions.

Lesson 1 • Backtesting Optimisation Strategies

Run walk-forward backtests of optimisation strategies on historical data. Validates whether theoretical gains survive realistic out-of-sample conditions.

Lesson 2 • Convex Optimisation with CVXPY

Model portfolio problems as disciplined convex programs using CVXPY syntax. Unlocks a broader class of objectives and constraints than SciPy alone.

Lesson 3 • Risk-Parity Portfolio Construction

Build equal-risk-contribution portfolios by equalising marginal risk shares. Provides a diversification alternative that avoids return estimation entirely.

Lesson 4 • Maximum Diversification and Other Objectives

Implement maximum diversification ratio and minimum CVaR objectives. Expands the toolkit for mandates that go beyond Sharpe ratio maximisation.

Lesson 1 • Optimising on BL Posterior Outputs

Feed posterior returns and covariance into MVO and CVXPY optimizers. Demonstrates how BL integrates with the optimisation frameworks built earlier.

Lesson 2 • BL Model Extensions and Critique

Explore factor-based views, entropy pooling, and known BL limitations. Prepares students to adapt the model to non-standard portfolio mandates.

Lesson 3 • Equilibrium Returns and Market Priors

Derive implied equilibrium returns from market capitalisation weights. Establishes the neutral starting point that Black-Litterman adjusts with views.

Lesson 4 • Posterior Return Computation

Combine prior and views using Bayesian updating to get posterior returns. Produces the blended return vector fed into the final optimizer.

Lesson 1 • Integrating ML Signals into Optimisation

Convert ML return forecasts into expected return inputs for the optimizer. Closes the loop between prediction models and portfolio weight generation.

Lesson 2 • Dimensionality Reduction Techniques

Apply PCA and autoencoders to compress high-dimensional return data. Reduces noise and multicollinearity before feeding data into portfolio models.

Lesson 3 • Clustering for Asset Grouping

Use k-means and hierarchical clustering to group assets by return behaviour. Supports diversification analysis and hierarchical portfolio construction.

Lesson 4 • Supervised Models for Return Forecasting

Train regularised regression and tree-based models to predict forward returns. Evaluates model skill using finance-specific metrics beyond standard R-squared.

Lesson 1 • Order Execution and Slippage Simulation

Simulate order execution with slippage models inside the backtest engine. Produces realistic performance estimates that account for market microstructure.

Lesson 2 • Transaction Cost Modeling

Model bid-ask spreads, market impact, and commissions in Python. Accurate cost modeling prevents overfitting to gross-return backtests.

Lesson 3 • Tax-Aware Portfolio Management

Incorporate tax-loss harvesting and holding-period rules into rebalancing logic. Improves after-tax returns without altering the pre-tax investment thesis.

Lesson 4 • Rebalancing Rules and Triggers

Compare calendar, threshold, and volatility-triggered rebalancing strategies. Choosing the right trigger directly affects net-of-cost portfolio performance.

Lesson 1 • Performance Measurement Framework

Compute return-based and risk-adjusted performance metrics systematically. Provides the quantitative vocabulary for evaluating any portfolio strategy.

Lesson 2 • Interactive Dashboards with Plotly and Dash

Build interactive portfolio dashboards using Plotly charts and Dash layouts. Enables stakeholders to explore portfolio data without writing code.

Lesson 3 • Regulatory and Client Reporting Standards

Align reporting outputs with performance presentation and disclosure standards. Ensures reports meet institutional and client-facing compliance requirements.

Lesson 4 • Automated Tearsheet Generation

Generate pyfolio-style tearsheets programmatically from a returns series. Standardises reporting and reduces manual effort in performance review cycles.