Seminar Description
Data Science and Econometrics for International Business
1 Overall Goal
This seminar equips second-year master’s students with the practical data science and econometric skills needed for applied research in international business. It builds directly on Theoretical and Empirical Research Methodology (Fall 2025), in which students learned R, Quarto, and basic linear regression.
The course follows a deliberate arc: from strengthening regression skills, through causal thinking, to modern data communication. The emphasis throughout is on intuition, real-world application, and reproducible workflow — not mathematical derivation.
2 Course Information
2.1 Logistics
- Programme: M.A. International Business Studies, Year 2
- Institution: Europa-Universität Flensburg
- Instructor: Prof. Dr. Claudius Gräbner-Radkowitsch
- Time: Thursdays, 16:00–19:00
- Period: 12 March – 18 June 2026
- In-person sessions: 9 (including recap)
2.2 Tools
- Language: R
- IDE: RStudio (via GitHub Codespaces)
- Reporting: Quarto (HTML and Word output)
- Key packages:
tidyverse,ggplot2,modelsummary,fixest,lmtest,sandwich,patchwork,plm - Bibliography: BibTeX / Zotero
3 Prior Knowledge
Students have completed Theoretical and Empirical Research Methodology (Fall 2025), which covered:
- R programming and RStudio (data wrangling, basic visualisation with
ggplot2) - Quarto for reproducible reports
- Foundational statistics: probability, sampling theory, hypothesis testing
- Simple and basic multiple linear regression (introduction only)
4 Git and GitHub
A key skill developed in this course is working with Git and GitHub — the version control and collaboration infrastructure that underlies virtually all professional data work, and that is increasingly essential in an environment where AI tools are integrated into everyday coding workflows.
Students are expected to create a free GitHub account before Session 1 and to register for the GitHub Student Developer Pack. Basic Git operations (pull, stage, commit, push) are introduced in Session 1 and used throughout the course. No prior experience is required; the GitHub Skills: Introduction to GitHub tutorial provides a self-contained introduction.
5 Pedagogical Approach
Each 3-hour in-person session follows roughly the following structure:
| Time | Activity |
|---|---|
| 60 min | Concept introduction: intuition, visuals, real examples |
| 30 min | Live coding demonstration |
| 60 min | Hands-on student exercise with a real dataset |
| 30 min | Debrief, interpretation, Q&A |
Core principles:
- Always lead with a business question
- Prioritise interpretation over mathematical derivation
- Use consistent datasets across multiple sessions where possible
- Embed Quarto communication skills progressively
6 Thematic Arc
The course follows a progression from description → association → causation → communication:
- Regression toolkit (Sessions 2–4): Deepen regression skills — multiple predictors, binary outcomes, diagnostics
- Causal thinking (Session 5): What separates a valid causal claim from a spurious correlation
- Panel methods (Session 6): Using longitudinal structure to control for unobservables
- Modern workflow (Sessions 7–8): AI-assisted coding and professional data communication
- Recap & outlook (Session 9): Consolidation and the inference vs. prediction distinction
7 Assessment
7.1 Final Exam
An open-book, on-site exam (details to be confirmed). Students may bring any materials but work individually on their own laptops.
7.2 Take-Home Tasks (mandatory, pass/fail)
Two structured exercises submitted as rendered Quarto HTML reports during instructor-unavailable periods. Pass criteria: all tasks attempted, code runs without errors, results are interpreted in plain language.
| Task | Dataset | Due |
|---|---|---|
| Task 1 — Earnings, age, and hours worked | cps-earnings | 02 April 2026, 23:59 |
| Task 2 — Hotel prices across European cities | hotels-europe | 29 April 2026, 23:59 |