Data Visualization in the Classroom

Today’s post is guest-authored by Alexander Von Hagen-Jamar, a postdoctoral researcher with the STANCE research program, in the  Department of Political Science at Lund University. His research and teaching focus on international relations, international security, state building and capacity, and empirical methodology.

In 2013, I spent two trimesters teaching at Carleton College in Northfield, Minnesota. While there, I had the opportunity to design a course about any subject I wished (within my expertise). I choose to organize the class topically around the consequences of violent political conflict. The other core learning goal was skill-oriented: I wanted to help the students develop applied quantitative literacy in context, and through doing so, encourage them to think deliberately about communication in a variety of mediums. To do that, I designed a series of assignments, centered around a group data visualization assignment.

The first assignment was to write a memo that summarized the military spending of a country over some period of time. The prompt was specific about what the memo needed – clear descriptions and several visual presentations of information (graphs, tables, etc) – but open about what country and what time-period. After that, the students worked in groups on the centerpiece assignment, a data visualization on a conflict-related topic of their choice. The summary description was: “Create a one-page tabloid-sized visual argument or description of some issue related to your group’s research question. That description must involve data (meaning between half and all of the visualization should summarize data in some way.) It need not fully answer the question, but it is important that it is connected to your question.” The presentation of the visual was bracketed by an in-class preliminary presentation of the research question, and an in-class reflective presentation of what worked and did not work. Finally, there was a second individual memo assignment, after the group project.

The students produced amazing data visualizations. They also worked very hard on them, and self-reported that they learned a lot. That, alone, was gratifying. I was concerned that the knowledge would not travel outside the assignment, though, so when I saw the results of the second memo assignments I was delighted and relieved. In the second memo, the visualizations were clearer, more purposeful (as in, they had definite communicative goals), more interesting, and riskier. The students tried more, and with more thought. By scaffolding the three assignments – two individual and one group – I was able to see learning that was (partially) out of context. It also gave us opportunities to talk about how data visualizations should differ in different contexts – the point of power point slide should be immediately clear, while a tabloid-style visualization benefits from, and can afford, richness that corresponds to the time the audience will spend with it.

Why did these assignments work? First and foremost, I was fortunate to have engaged students who wanted to be there – getting buy-in from political science majors in a mid-level, optional political science course is comparatively easy. I did a lot of selling anyway. Second, I was able to give the students the time and resources they needed to work on the assignment. In particular, I was able to give them class time to work on it, and even more importantly, to bring in experts to help the students with the necessary technology. Carleton has fantastic Academic Technology staff who did a lot of the heavy lifting on the technical instruction – the group assignment only worked with their help. Finally, making the most difficult and unusual assignment a group assignment helped, as the students were able (required) to help each other through the most daunting aspects.

I would do these collection of assignments again, but only if the conditions I described above were met, or I had substantial substitutes for them. When I did a “lite” version of the assignment in a required methods course at the same institution, the results were underwhelming: I managed to stress the students out, without, I think, teaching them very much about presentation, quantitative literacy, or the topics of their interest. But given those pre-existing conditions: buy-in from the students, substantial support from the institution in terms of staff time and expertise, and assessment flexibility that allows major group projects, I think scaffolded assignments like this are worthwhile and interesting ways promote quantitative literacy and subject-level expertise in mid-level and advanced undergraduates.