How To Read This Chart: Actually, maybe do look up

Saturday, January 29, 2022

Read online

By Philip Bump

Did King Ghidorah forward this to you? Sign up here.

If you look up right now, what do you see?

I'm guessing probably a ceiling; most of our readers are in the Northern Hemisphere (I assume, just based on, well, everything), where it is winter, and that means a greater chance of being inside than not. But if you go outside, what do you see when you look up?

The answer to that question depends on a few things. It depends on the time of day and on the weather and on the season. Right now, as I write, if I went outside and looked up I would just see an expanse of blue before I died of hypothermia. It depends.

Enter Cees Bassa. Bassa is an astronomer in the Netherlands who uses radio telescopes to study pulsars. And, like so many of us, he also takes his work home with him. So he cobbled together a little camera with a fish-eye lens attached to a small Raspberry Pi computer and, as he explained when we spoke on Wednesday, put it on the roof. Its task? Take a photograph of the sky every 15 seconds for a year.

And so it did.

From the Twitter feed of Cees Bassa. Used with permission.

This is just such a terrific bit of data visualization that I find it hard to contain myself. There's so much to be gleaned from it!

First, though, some orientation. Each strip represents one day of 2021, from January at the top to December at the bottom. Each strip begins at 4 p.m. and ends 24 hours later. It progresses in 15-second increments, though those are hard to pick out at this resolution.

So what can we learn? An enormous amount.

The first thing you'll notice is the hourglass shape of each night, from long nights in winter to those remarkably short nights near the summer solstice. The nights are particularly short during the summer at Bassa's house because he lives fairly far north, at about the same latitude as Edmonton.
Then, notice the white stripes across the nighttime hours. That's the moon, progressing across the sky 12 times in 2021.
Now, notice that the edges of the daylight periods next to the hourglass are more blue. That, Bassa explained, is because the Sun has just set or is about to dawn, lightening the sky but not illuminating clouds.
The bright blur to the right of the image is the Sun itself, high in the sky at midday though more visible in the summer months.
You can also see the two periods where the camera lost power, but we need not dwell on those.

In the larger photos obtained by the camera, Bassa — obviously a trained eye — was able to pick out the movement of constellations and meteors streaking by overhead. If you'd like to download and explore the entire 378 megabyte photo, he's made it publicly available.

Nerdery + technology + education + data visualization? We should all strive for such a combination.

What is going on here?

(At some point I will probably stop explaining what the title of this section means, which is that I'm about to clown on something, but that point has not yet arrived. So.)

One of the benefits of the Internet is that it brings the world together, allowing, for example, for a reporter in the United States to see a complicated visualization of the sky over an astronomer's house in the Netherlands.

One of the drawbacks to the Internet is that it allows media organizations to bring people together to respond to dumb, unscientific surveys, the results of which are then presented as useful, which they aren't.

This newsletter will now progress from describing the first phenomenon to the second.

Consider this survey, conducted by LinkedIn apparently as one of those 𝓮𝓷𝓰𝓪𝓰𝓮𝓶𝓮𝓷𝓽 ploys of which brands are so fond.

There's nothing useful about this. It's a public survey to which anyone can respond meaning that, should someone have had so desolate a life, they could rig the results however they wanted. It's like an ice cream parlor that turns its tip jars into a mock presidential election; the point isn't to figure out who's going to win, it's to get more tips.

And yet, there was that LinkedIn survey, mentioned uncritically in a local newspaper in New York.

"In a recent poll of more than 21,000 LinkedIn members," Emma Goldberg wrote for the New York Times, "59 percent said a colleague's departure had led them to consider quitting as well."

And in a poll I conducted of my Twitter followers in October 2020, 41 percent of people responded that, yes, Superman was real. LinkedIn got a few more responses, but its results were no less dubious.

Meanwhile, the Boston Globe was up to the same sort of thing. In a column about threats to elected officials — certainly a legitimate point of concern! — reference is made to "a Boston.com online survey" in which "over 90 percent of respondents felt there is nothing wrong with anti-vax protests" at the mayor's home.

That survey looked like this.

Not only is this not scientific, it's precisely the sort of thing people would game! It's a question about a potent political topic that also provides an opportunity to cast something seen as toxic as socially acceptable.

I understand that this is a newsletter about charts, and I will not try to pretend that the salience of that LinkedIn poll is based on it including a bar chart though technically it does and therefore technically I am fully within my rights to include it here. (Okay, so I will try to pretend that, I guess.) Instead, the point is that the more important part of data visualization is data, that while you can paint a very elegant and appealing picture of horse manure, it's still horse manure.

So please don't go out of your way to generate horse manure — and certainly don't pretend that horse manure is not horse manure.

How SCOTUS got where it is

This is the section of the newsletter where I take a nice chart that I made and make you look at it.

This one I put together after news broke of Justice Stephen G. Breyer's retirement from the Supreme Court. I like to think about those seats and Senate seats in terms of historic continuity, so I grabbed 160 years of data from the Senate historian's website and figured out whose tenures flowed into the ones we see today. Time passes from top to bottom, with years indicated at left.

(The line that crosses over is William Rehnquist's elevation to chief justice.)

The point was to show how the confirmation processes for justices went from shrugging voice votes to near-unanimous recorded votes to brawls that resulted in near-ties over and over again. The point was also to make something nice to look at, like a beautiful painting of horse manure.

I will also note that, after last week's newsletter, I got good feedback about the lack of contrast between the colors in one of the charts. So I've been more attentive to that this week, proof that I, unlike a donkey or a chair, am able to learn and grow.