Teaching Viz Digest No. 1: Exercise is Good for You

It’s been just a little over one week since #teaching-vis has launched on the Data Visualisation Society (DVS) Slack. And it’s quite something.

We’re a mixed bunch that includes academics, consultants, researchers, designers, and dataviz enthusiasts.

We teach students, co-workers, organisations, scientists, journalists, designers, the general public and some of us even teach our friends.

And, following in the ethos of DVS, it’s inspiring to hear discussion among D3 wizards, Tableau Jedi, R gurus and pen & paper geniuses.

But more than all of that, it’s amazing to share ideas about how we teach data visualisation.

A big topic in this first week has been around exercises we use to teach our students different aspects of dataviz.

This began with a great question from Enrico Bertini on whether the group had ideas for exercises that teach students how to use colour in visualisations. This discussion quickly snowballed into a document where we’re adding some of our favourite exercises (and not just on colour).

So without further ado (and there’s been a lot ado, seeing it’s issue one of the digest), here are some the exercises that have been shared.

Cutting up the Visual Vocabulary

Having taught the Visual Vocabulary in the post-lunch graveyard slot by talking through the charts, I was delighted to see this exercise submitted by Henry Lau and originally developed by Martin Chorley.

Asking groups to re-create the FT’s brilliant Visual Vocabulary from cut-outs sounds like an ingenious way to get people to think out loud about choosing the right chart for their data.

Turning the vocabulary into a puzzle and making students piece it together sounds like a neat trick to help expand their use of common chart types.


Making awful graphs

This is one I’ve used with students to try to get them to bring together a few ideas around design. Students are asked to create a really horrible graph. But the catch is, they have to have a clear rationale for every design decision they make.

Confusing colours. Check. Nonsensical scales. Check. Hand-made chart-junk. Check. Double-perspective 3D… wait, what?

It might seem counter-intuitive, but, what my students have referred to as “chart-therapy”, gives them an opportunity to think through design decisions and vocalise them.

It’s a safe way of talking about things they might have done in the past.

And it also happens to be a lot of fun.

Just make sure to emphasise this is what your students shouldn’t do.


Estimating values

Martin Chorley also offered up this very interesting exercise that helps his students think about how chart choices influence the readability of the data they represent.

Using the same dataset, he offers students a set of different charts, but drops the axes and labels. One of these might be a standard bar chart. Another might be a treemap. And a third might be a 3D pie chart.

A quick example from Martin’s exercise.

Students are asked to guess the values represented in each, and are in for a surprise when they realise that they’re looking at the same data throughout.

It sounds like a great way to start a discussion on how the choices we make in representing data can lead to different interpretations.

As a bonus, Martin suggests also capturing the estimates the students make, to show which charts lead to the best guesses.


But what about colour?

Homage to the square Blue Secret II by Josef Albers. Photograph by Chrisi1964, distributed under a CC-BY 2.0 license.

Going back to Enrico’s original question there were a few suggestions.

Annette Greiner mentioned using Josef Albers’s colour squares as a way to get her students to think about colour relationship and especially how colour is relative.

Martin Chorley offered an exercise that asks students to map out implicit and explicit semantic associations with colours and colour combinations. It gets students thinking about accessibility and colour in a cultural context.

I suggested using a colouring-book approach that asks students to use colour to emphasise specific data relationships in a multi-chart display.

If you’re interested in finding out more, join us over on the #teaching-vis channel on Data Visualisation Society.

And if you’d like to get involved in building out our ultimate exercise list, you can do so by going to the Google Doc.

A huge thank you to all participants in the #teaching-vis channel and especially Martin Chorley, Enrico Bertini, Henry Lau and Annette Greiner.