Session 9: Recap and Looking Ahead
Note
Date: Thursday, 18 June 2026, 16:00–19:00
Learning goals
- Consolidate key concepts from the full course arc (description → association → causation → communication)
- Reflect on the inference vs. prediction distinction: when do we estimate to understand, and when to forecast?
- Discuss what comes next: an introduction to machine learning thinking and where to go from here
- Prepare for the final exam: what to expect and how to approach it
Course arc recap
| Sessions | Theme |
|---|---|
| 1–4 | Regression toolkit: multiple predictors, binary outcomes, diagnostics |
| 5 | Causal thinking: confounding, selection bias, the identification problem |
| 6 | Panel data: fixed effects and unobserved heterogeneity |
| 7–8 | Modern workflow: AI-assisted coding and data communication |
| 9 | Consolidation: inference vs. prediction, looking ahead |
The inference vs. prediction distinction
One of the most important conceptual lines in applied data science: are we estimating a coefficient to understand a relationship, or are we building a model to make forecasts? These require different choices — in model complexity, evaluation criteria, and how we report results. This session addresses the distinction conceptually, without a dedicated technical session on machine learning.
Important
Material will be available here before the session.