Show Your Stripes

About a year ago, I ran across the cool new site for the #ShowYourStripes campaign.

On the site, you can download a customized climate anomaly graph for any country or state. Spoiler alert: they’re all trending from blue towards red as the global climate warms, but there is quite a bit of variability from place to place.

These stripes are a simple but effective way to help emphasize the global warming trends we’ve seen over the last century, and how they’ve impacted practically every corner of the globe.

Warming Stripes for <All of USA srcset= from 1895-2019″>

Last year, there was a social media campaign for the site around the solstice, and it looks like it happened again this year.

Because the the images are red, white and blue, I think they can also be used as a fun (if concerning) way to recognize Independence Day. Given the color order, they might even be more appropriate for Bastille Day, which is also coming up. But I digress.

Data in Context

After looking at a few of the graphs individually, I immediately began to wonder how locations compared with each other.

In particular, because it’s most relevant to me, I wondered how New Jersey compared to global average.

While you cannot easily compare locations on the site, it didn’t take too much effort to download the two graphs and mash them up in Photoshop.

Note, if you do this yourself, you may have to align the graphs manually, because the time range available for each graph depends on the underlying dataset. For example, the global data set is available from 1850 to 2019, while New Jersey only has data available from 1895 to 2019. In addition, I believe the colors are not an exact match either, as they appear to be scaled using a baseline average specific to each individual dataset, as well as the localized standard deviation. (This is common when calculating anomalies.)

In the end, the overall trend is clear for pretty much for any location you choose. But when you compare the graphs, you also see quite a bit of inter-annual variability between the global average and a specific location.

Ultimately, that’s what struck me about the New Jersey graph – there is more year-to-year variability in our localized area, compared with the smoother overall global trend. For example, there are years where the local anomaly is much higher than the global anomaly, and years where the local average is actually negative even when the global average is well above.

Which made me wonder, perhaps some of the pushback against the scientific consensus of gradually increasing temperatures, is because that message doesn’t jive with people’s lived experience?

By definition, those experiences are local.

For me, there are two takeaways here.

In visualizing a dataset, it is important to provide context. That includes making sure the data is understandable, for example, by providing min/max ranges or making the trend clear.
More importantly, we need to make sure visualizations are relatable, that is, relevant to an individual viewer. This requires putting data in a context that a person can understand based on their own personal experience, especially for those who are not as familiar with the science or visualization format.

One of the key advantages of the ShowYourStripes site is that it allows you to see graphs for the global trend, as well as those for more local areas like your country or state. And it is those more local graphs that you can interpret and understand based on your own experiences.

So, as science communicators (and educators), if we want to make a stronger message (or have our lessons understood), we need to make sure our data is relevant to others by aligning our visualizations with a viewer’s personal knowledge and experience.

Yes, I know this is far easier said than done, but we have to keep trying, and the Show Your Stripes campaign is a great example of how this can be accomplished. (I just wish they allowed you to create a stripe with both global and local data for easy comparison.)