International collaboration in teaching topological data analysis
The project
The main part was a project based course where students from 4 universities (ETH, TU Eindhoven, St. Louis University, De Paul University Chicago) collaborated for one semester on open problems in the area of topological data analysis and its applications. In this first iteration we had three groups of 3 students each, and each group was advised by 2 mentors. The underlying idea was that the course trains the students through project-base learning that allows them to apply their newly acquired theoretical skill to a concrete problem, as well as through strengthening the transferable skill of group work in an international online setting.
Implementation into teaching practice
As the goal was the implementation of a new course, the project was directly implemented in teaching. The main steps of the process were finding project partners, developing interesting problems for the groups to work on and, of course, mentoring the student working groups.
Lessons learned and further impacts
We consider the most important goals of the project to be achieved: the students have gained experience in applying topological data analysis and have expanded their international network. One of the groups is planning to publish their findings of the project. Another student from ETH plans to continue her work as a thesis, which will be co-supervised by the mentors from Eindhoven and St. Louis. Both the international collaboration aspect as well as the project-based learning were very interesting both for the students as well as for the mentors, as a survey among all participants has shown. One issue that we want to address in future iterations of the course are methods to ensure that the workload is distributed more evenly within the groups. The general concept of project-based learning in an international setting should translate nicely into other contexts. However, we noticed that close mentoring of the groups is an important factor for the success of the project, which means that such approaches are less suitable for larger student populations.