Interactive open-code computer simulation exercises promote a deeper understanding of quantitative phenomena in bioanalytics

We have designed interactive, open-code simulations to enable student exploration of non-intuitive quantitative phenomena in bioanalytics. The simulations have been built on the LET’s JupyterHub platform. With this project, the DBIOL’s Center for Active Learning (CAL) has acquired the competence to develop such teaching tools in-house to boost the computational reasoning skills of our students.

We developed around a dozen JupyterNotebooks that allow students to explore quantitative aspects involved in various bioanalytical techniques as part of the selfstudy component of the course. Besides the actual computational activities, these notebooks contain background information, references to the relevant course material and detailed instructions. These notebooks were made available to students via the course’s Moodle page with the ETH JupyterHub serving as the computational back end. Jupyter Notebooks were introduced in a 15 min session at the end of the course’s first lecture and a 2 hour session was offered on a lecture-free day about half way through the course.

The project had two main goals. One goal was to develop competence for the creation and deployment of JupyterNotebooks on the ETH JupyterHub as teaching tools in the DBIOL. This goal has been fully achieved. We have developed and successfully deployed several such notebooks including a broad range of technical features. We have also identified a pool of teaching assistants, who are able to help with this development. Finally, we were able to convince the organisers of another of the department’s course to switch their computation exercises from Matlab to JupterNotebooks.

The other goal was to help students gain a better understanding of some of the less intuitive quantitative aspects of bioanalytics by exploring these aspects with the help of interactive JupyterNotebooks. This second aspect turned out to be more challenging than expected. Based on the log data, discussions with the semester speakers and a focus group of students from the course, most students saw the JupyterNotebooks as «cool» but «not exam relevant». Also, the notebooks combined several activities into a single notebook, which made working through a single notebook a substantial time investment. In a semester, that already has a relatively high workload, this led to students directing their attention elsewhere.
To increase student engagement, we are now working on three aspects: 1) With the upcoming revision of the course, we plan to include a 2 hour in class session early in the Semester that introduces Students to JupyterNotebooks in general as well as the discussion of Notebooks related to problems from the previous lecture in this course. 2) We have convinced one of the course’s lecturers to transform one of her on-paper in-class exercises into in-class JupyterNotebook-based exercises. We are hoping to convince other lecturers as well. 3) We will split up some of the longer notebooks into a series of shorter notebooks. Where each of these shorter notebooks focuses on a single concept. This way students can complete these individual notebooks in a reasonable amount of time and select those notebooks that are particularly relevant for them.