Toolbox Courses in Data Science
This project addresses the didactical challenges of toolbox courses in data science, which serve students from multiple programmes with diverse needs.
In the literature on toolbox courses, a recurrent observation is the difficulty of ensuring visible and functional connection between the courses and study programme they “serve”. For financial and other reasons, courses often have students from several programmes, and the experts of data science who teach them do not have updated and detailed knowledge of the needs for data science elements in later courses within each of the programmes. A main objective of the project is therefore to develop methodologies to identify and address such needs.
The aims are twofold:
- identify generalized data science needs by studying paths of reasoning from selected situations and questions in undergraduate science to data science elements, such as a particular testing method from inferential statistics,
- Then, to design and experiment didactical situations for data science toolbox courses focusing on those needs, in view of preparing students for deploying data science tools in their later studies.
WP Leader: Professor Carl Winsløw, UCPH
Researcher: PhD student Jonas Rysgaard Jensen, UCPH
Researcher: Postdoc, UCPH
Four science programs, two at AU and two at KU, along with, for each program, a data science toolbox course will be selected for in-depth analysis. Central to this project is the development of praxeological (reference) models (PRMs) that can be used in the analysis of data science needs in specific programs. This work will inherently be an iterative process, where the models are continuously refined and revised as they are applied to or confronted with new data.
Initial input for the praxeological analysis will in each case be analysis of course curricula for the selected courses, overall curricula for the relevant program, interviews with directors of the program, and guidelines and recommendations from relevant international organizations, domain specific as well as data science/statistics organizations (e.g. ASA’s GAISE report, RSS’ suggestions for data science courses, or ICH’s guidelines on the statistical analysis of clinical trials). To go beyond merely declared needs and goals, important further input to and refinement of the models will be based on interviews with teachers, students, analysis of actual teaching materials, as well as analysis of select student work.
WP Leader: Professor Carl Winsløw, UCPH
Researcher: PhD student Jonas Rysgaard Jensen, UCPH
Researcher: Postdoc, UCPH
This work package will be carried out by a postdoc with a PhD in data science education at the university level, including a strong background in applied statistics and computing. The three-year postdoc project will follow the methodology of didactical engineering[1] (a specific kind of “design research” based on ATD, with origins back to the early 1980s) but with highly innovative content, enabled by the results of WP1.
[1] See: Barquero, B., Bosch, M. (2015). Didactic Engineering as a Research Methodology: From Fundamental Situations to Study and Research Paths. In: Watson, A., Ohtani, M. (eds) Task Design In Mathematics Education. New ICMI Study Series. Springer, Cham.
.jpg)