Sparklines: Comparable or not?

When I was a child, my brother found a Super Nintendo next to some garbage bin. It came with four games. And it was fully functional. My brother and I joked that some parent probably had decided that gaming might be not the best use of time for their kids. So now we got to spend our time with it. And boy – it was fun! I remember lots of hot summer days laying in front of a very old TV, trying to break my own record of Yoshi’s Island levels.

The Super Nintendo was the only console I’ve ever owned. But when I think back to it, I smile – and I’m sure consoles put a lot of smiles on a lot of faces.

This week, I once again used our fresh table tool – and this time with sparklines; these tiny line charts in the last column. Thanks to them, we can see that the Wii was the most successful console of the five consoles shown here. In fact, the Wii was one of only three consoles that were sold more than 100 million times. The other two ones were the PlayStation and the PlayStation 2.

Chart choices

Sparklines are curious things. They’re supposed to show a trend, and a trend only. They’re supposed to show when something (like stocks) increase and decrease, where the peaks and the valleys are. But sparklines are not supposed to be comparable with each other.

So when you’re seeing two sparklines with the same height, the ebbs and flows of the first one could play out between 0 and 10 (e.g. US-Dollar), while the other sparkline’s peak is at 10,000.

But that’s odd, no? Doesn’t that invite people to make totally false assumptions? The answer is: Yes. Likely so. But sparklines need to get comfortable in a tiny space anyway – if you set the same y-axis range for two sparklines with different dimensions, you’ll end up with a visually flat line for one of them. Which goes against the whole idea of using sparklines to show trends.

But there are ways to increase the chances that your readers don’t misread the visualization. If you create sparklines with different y-axis ranges, label the first and last value (that’s a default in our table tool). And if your y-axis numbers are similar anyway, choose the same y-axis range; as I did in the table above.

I’ll be on vacation for the next two weeks – which means that you’ll be able to hear some other voices! Our developers Simon and Fabian agreed to write the next two Weekly Charts, and I hope you’re as curious as I am to see what they’ll end up talking about. Until next week!