useR! 2024

In my talk, I explore teaching data science and causal inference in a mandatory undergraduate economics course using R, with no prior student knowledge in programming, causal inference, or project management. To teach these concepts, I opted for a flipped classroom approach and group projects: Initially, students learn theoretical concepts (descriptive statistics, data visualization, and causal inference) through short videos, foster these concepts through individual RTutor Problem Sets, and apply them in a case study, which we discuss in class, using RMarkdown to make the results reproducible. Subsequently, they engage in groups of three students to solve three economic projects in RMarkdown, generating HTML reports. These projects require them to gather data (e.g. via API or web scraping), perform statistical analyses, and infer causal relationships, collaborating on private GitHub repositories. Peer reviews follow each project. Tools used include the R-packages RTutor for Problem Set design, ghclass for GitHub management, and GitHub Actions for report verification.