Data viz as a teaching tool: some experiments

A while ago an editor at Mondadori Education contacted me because they wanted something to try something new and fun for their science high school textbooks. So they asked me to create an infographic for each chapter. But the thing is:

what if we used them as an exercise as well, instead of just conveying information to the students?

You see, most of the visualizations that we create — and read — are in a way “passive”. They tell the reader: “Hey! Look at this pretty bubbles and lines: there’s something to learn here if you read me”.

Yet at times you might want something “active” — an object on which students can put their hands on. And with a little luck that should make it more enjoyable than a normal visualization. I guess you might call it an interactive, because in a way it is, but it differs from the usual interactives in some fundamental ways:

  • It doesn’t have to include all data that the reader needs to know;
  • It shifts some of the burden from the author to the reader (lazy people: hooray!);
  • It might be even wrong (?).

For the sake of brevity, we just called them “active infographics”. (We don’t get called creative people for nothing after all.)

So I made a bunch of them, and I wanted to share here some drafts to see what everyone thinks, and hopefully understand how to make them better. But enough with the talking, let me show a few examples (as usual they are in Italian, but bear with me).

Two caveats: we didn’t want to get too fancy with exotic visuals — they’re high schoolers not PhDs — and those examples are all drafts, so it’s possible they still contain some — unintentional, we’ll see why this distinction matters in a bit — errors that have been corrected by the experts later in the process.

Each visualization is composed of the graphics, a brief text that explains what kind of chart are we using and how to read it, and a few questions that have to be answered by the students. For that, they’re going to have to use what they see in the visualization, what they’ve learned in the chapter, and most importantly their reasoning abilities.

That said, here we go:

In the first exercise, I̶n̶d̶i̶a̶n̶a̶ ̶J̶o̶n̶e̶s̶ students have found four ancient fossils and has to determine how old they are. To do that, they may use two different tecniques, i.e. radiocarbon dating or K-R (potassium-argon) dating. Both are based on the relative presence of some radioactive elements in the fossil, so the age of each one can be deduced by the atoms at the bottom.

After that, they just have to indicate the correct position of the fossil near the timeline at the center. The scale had to be logarithmic, which is not ideal, but we have to go back both billions and thousands of years ago so I guess there was no real alternative.

Here students don’t have to do much in terms of drawing, but they really do have to know how the different methods work (the two charts at each side give also a small recap, just in case).

This one is pretty simple: a timeline of earthquakes in Italy, Japan and United States on the x, with the magnitude on the y. The size of the circle is proportional to the number of victims. There are a few events missing, like the recent 2016 earthquake that struck the center of Italy, and students have to draw the relative bubble with the correct attributes.

The circles one the bottom are there to give them a rough measure for size. As fun as trigonometry is (so to speak), we wanted them to spend most of their time understanding the events and not on radiuses (radii? Radiuseseses? English, behave!) and areas.

The third one is about Italian volcanoes, and it asks students to determine which and where they are, how tall they are, and when did they last erupt. Some of them pose an interesting challenge, as they are partly underwater like the Stromboli.

Since I’m (proudly) a terrible person I didn’t bother telling students any of this — I gave them only visual clues so they mostly have to figure that out by themselves.

This one is about one of the most fascinating topics of all times [̶s̶e̶x̶ ̶–̶ ̶s̶o̶r̶r̶y̶,̶ ̶w̶e̶ ̶c̶a̶n̶’̶t̶ ̶t̶a̶l̶k̶ ̶a̶b̶o̶u̶t̶ ̶t̶h̶a̶t̶ ̶i̶n̶ ̶h̶i̶g̶h̶ ̶s̶c̶h̶o̶o̶l̶.̶ ̶A̶n̶y̶b̶o̶d̶y̶ ̶k̶n̶o̶w̶s̶ ̶t̶e̶e̶n̶a̶g̶e̶r̶s̶ ̶h̶a̶v̶e̶ ̶n̶o̶t̶h̶i̶n̶g̶ ̶t̶o̶ ̶l̶e̶a̶r̶n̶.̶ ̶N̶o̶t̶ ̶t̶h̶a̶t̶ ̶t̶h̶e̶y̶’̶r̶e̶ ̶i̶n̶t̶e̶r̶e̶s̶t̶e̶d̶ ̶a̶n̶y̶w̶a̶y̶]̶ geology. Here I designed a few imaginary tectonic plaques, each with different characteristics. Students have to figure out what will happen when they move around at different speeds (spoiler: mostly bad, bad things).

In my first sketches I tried doing this with the real plaques, but very soon it became impossible to visualize them in any comprehensible way. You start with a 3D space that you have to flatten (which is not as fun for the rest of humanity as is for us), and on top of that you have to add moving surfaces, and they all have different sizes and borders, and it’s not that one border is like another — there are several types and it matters when they get on top of each other because reasons.

Long story short, you end up in the kitchen crying and the ice cream stash gone, which is not great.

But then, luckily, there’s one of the few design principles that is true maybe 99% of the time (a world record in our field, I believe): DO NOT USE 3D CHARTS. So I threw everything away, made a much simpler 2D model and my editor liked it instantly.

Lastly, we’ve got a scatter plot that represents planets in the Solar System. On the x there is the average density, on the y the average distance from the Sun. Sizes are proportional to the real ones, of course, while the color shows the most common element in the planet atmosphere.

Here I purposely included several mistakes and told the students to correct them. (The journalistic side of me got a little I’m-the-bad-guy-now kick off of that, gotta admit it. And with what we have been seeing lately, I assume adding “fake-yet-believable-chart scientist” to my résumé will get me pretty far career wise.

You hear that, Donald?

But I also think there’s an important lesson to be learned here: never trust information just because it’s been given to you: be skeptic, try to understand if it makes sense and is compatible with other evidence).

Anyways, density of the Earth is most definitely wrong, as are distance and color of Uranus [̶i̶n̶s̶e̶r̶t̶ ̶T̶r̶u̶m̶p̶/̶g̶o̶l̶d̶e̶n̶ ̶s̶h̶o̶w̶e̶r̶ ̶j̶o̶k̶e̶ ̶h̶e̶r̶e̶]̶. I also added some visual cues to make it a little easier to identify the correct spots. Still, the point of the exercise is not the be as precise as possible — machines do that pretty well, thank you — but to understand why each element belongs where it belongs.

So those are a few examples. But I am curious: do you know if anyone else is doing something similar in the field?

(I’m sure they do: I mean, Scott Klein’s great newsletter Above Chart routinely comes up with the coolest data journalism from Ancient Egypt or something, apparently to embarrass folks that think they’re doing something new, so I have no doubt there will be plenty of others in this vein as well. But please do post some links, I’m interesting in seeing what they have been doing and how can this be improved. And even if you don’t have any, comments, critiques and suggestions are of course welcome.)

I am also wondering if it wouldn’t be nice to put out something similar in news outlets as well. I do recall an experiment in the New York Times, a while ago, in which they asked readers to draw something inside an interactive (I couldn’t be even more vague if I tried, but I’m and old man and my memory is what it is), share it and compare results, but not much other than that.

I know it’d be a little of a burden on the newsroom in terms of coding, but you never know. Might this be something like a weekly quiz or something? Just throwing the idea around.

Lastly, since you’ve been so kind getting this far — and let’s be honest, no one but a few data nerds will ever read this — I can reveal you a little secret. This was a plot all along. The real purpose of all this endeavor is entirely another.

We always complain that people can’t read our works properly, that visualizations are great and everyone should have more. That seven-and-a-half page spread in the last issue? Come on, that was barely enough for my charts! Visual literacy anybody? So why not get people addicted when they’re still young and learning, is the plan. More fun for them, more work for us. Right?

(Ok, this bad guy thing has now got officially out of control.)