Course Materials
Presentation & objectives
These two complementary lessons introduce the concepts needed to understand the practical case studies. The first develops the foundations for selecting and evaluating machine learning methods; the second explains how exploratory work evolves into maintainable Python pipelines.
By the end of these lessons, you will be able to:
- compare supervised and unsupervised learning methods;
- select appropriate validation, tuning, and evaluation strategies;
- interpret model results and statistical evidence carefully;
- identify the limitations of notebook-based workflows;
- explain how modularity, testing, and separation of concerns support reliable ML systems.
Lessons
-
Machine Learning Techniques: From Theory to Practice — This one-hour lesson develops the conceptual foundation for distinguishing learning problems, comparing common algorithms, applying cross-validation and hyperparameter tuning, selecting evaluation metrics, and interpreting p-values without confusing statistical evidence with effect size.
-
Towards Production-Ready ML — This 30-minute lesson prepares the Air Quality activities by explaining notebook reproducibility issues, modularity, separation of concerns, dependency injection, interface-based design, and the respective roles of unit, integration, end-to-end, and performance tests.