In this seminar students will learn about modern statistical workflows and apply them to statistical forecasting problems. As part of the course, students will learn how to set up analysis workflows in R with particular emphasis on automation and reproducibility. They will then use these workflows to design and evaluate forecasting systems. In addition to the more hands-on workflow component, the methodological focus is on diagnostic assessment and uncertainty quantification of forecasts. To regularly evaluate their progress throughout the seminar students will use their systems for forecasting challenges in different contexts. Intended Audience: students in (bio)mathematics or other programs with a strong statistics or data science component or students who have a strong interest in quantitative methods and statistical modeling. Requirements: Lecture in statistics, e.g. B.Sc. lecture Statistik or an equivalent applied statistics lecture, and at least basic knowledge of R. Possibly helpful: lecture in machine learning. |
