Blog: How do we measure change?

How do we measure change over time?

The y-axis can be so deceptive. Change the scale a little, and a small change looks massive and overwhelming. It can change the entire story of the graph.

…but also I encourage students to think about the story told by the shape of the data first, and then we attend to the scale to adjust the story. This plays out beautifully in the slow reveal with a data viz by Bo McCready (@BoKnowsData on socials): Disney Princess Baby Names in the US (1951 – 2021)

Here’s the full data visualization:

There’s a ton of data in there! You might need to click to get a closer look.

The graph shows the “annual prevalence of each [Disney Princess] name in the USA. The gray vertical lines show when the first film including a princess with that name was released” (McCready, 2022). There are some princess names that seem to be greatly impacted by Disney’s influenced, or the “Disney bump.” Check out how this looks for Merida, the heroine in 2012’s Brave.

Very few babies in the US were named Merida, and then… SLAM. It’s like a wall of new Merida babes.

Meanwhile, Aurora experienced a surge, but it’s hard to say it pin that on the influence of Sleeping Beauty, released in 1959. However, there was a lot of new marketing around the Disney princesses that may correlate with the increase in Auroras across the US.

And then there are some names that experienced a short-lived Disney bump, like Tiana, the star of 2009’s The Princess and the Frog.

So it would seem like Merida experienced the biggest increase from the Disney bump! It looks like an extra… 90? According to SSA data, 110 babies in 2013 were given the name Merida, compared with 19 in 2012. That’s a huge percentage increase (579%!), but this maybe explains why I have yet to meet a child named Merida, even with the dramatic shape of the data.

Meanwhile, Tiana was the given name of 478 US babies born in the US in 2008, and 969 babies born in the US in 2010. The rate about doubled after the movie The Princess and the Frog premiered, but that resulted in a net increase of about 500 babies per year. You’re much more likely to encounter a Tiana than a Merida.

It’s more challenging to compare graphs that use different scales along the y-axis.

One advantage of using the slow reveal graph routine, is that we can focus on different features of the graph, and revise our thinking.

In the slow reveal that I created, I revealed only two y-axis at first. They both happened to feature the same scale. What do those numbers mean: are they percentages? Or total babies with that name in a given year? (spoiler: it’s the latter)

Will the other princess names feature the same scale along the y-axis? I know so many students named Anna, and I’ve never personally had a student named Mulan. So it seems like the scale might be different. (And, of course, it is.)

From there, we could elicit a conversation about how we measure change. If we want to answer the slightly ambiguous question about which name’s popularity was most influenced by its Disney movie, we have to decide how we measure change. Is it percentage increase? Is it the increase in the number of babies given that name the year after the release of the film? How can we use the graphs to calculate these two separate measures?

The slow reveal graph for Bo McCready’s Disney Princess graph is available on SlowRevealGraphs.com now. After playing around with this data visualization, students might enjoy looking at the inverse effect sales of Amazon’s Alexa had on the name Alexa.