More Adventures in Terrible Data Sonification

Data sonification is a touchy issue at Mad Art Lab. I’ll admit, when I first started writing here, I was on board with all of it — stock prices? Twitter conversations? The Higgs-Boson? If you turn it into music, no matter what it sounds like, I guarantee I’ll be impressed.

But an article musician Sarah Angliss wrote when she was a Lab contributor changed my mind on the matter. Her point was that good data sonification has to reveal something about the data that couldn’t be revealed any other way; it has to succeed where the charts and graphs and scientific lingo failed. It’s a hard thing to achieve, which is why it’s not surprising that a lot of the stuff the media gloms onto doesn’t achieve it. Most of it simply sounds like listening to a page full of numbers, rather than reading one.

Which is precisely where A Song of Our Warming Planet lands. Through the urging of his geography professor, University of Minnesota undergrad Daniel Crawford decided to take surface temperature data from NASA’s Goddard Institute of Space Studies from 1880 to the early 2000s and put it to music. It appears to be a great idea. Because of the ease with which we hear pitches raise and lower, music seems to be the perfect vehicle with which to demonstrate how temperatures have increased over the last century.

According to Ensia, here’s how he worked it out:

The temperature data were mapped over a range of three octaves, with the coldest year on record (1909) set to the lowest note on the cello (open C). Each ascending halftone is equal to roughly 0.03°C of planetary warming.

In Crawford’s composition, each note represents a year, ordered from 1880 to 2012. The pitch reflects the average temperature of the planet relative to the 1951–80 base line. Low notes represent relatively cool years, while high notes signify relatively warm ones.

Every 0.03°C is a half step, every year is a quarter note. Sounds good. Then what?

Then nothing. That’s where he stopped. The piece is a series of seemingly random quarter notes. True, the line does start to veer higher near the end, but there’s no emotion or inspiration or moment of sudden understanding to the thing.

I’m not going to beat up on an undergrad for trying something he found interesting. I would have been so excited to write something like this at all in college, much less have it make io9 and go all viral.

I will beat up on the media a bit, though. It’s not just because they’re calling something ingenius that’s been done in this exact boring way a thousand times — I’ve already made my points about that. It’s because they missed a key bit of information that changes this from a mundane scale study into a million interesting projects in the future. That information is this:

Crawford hopes other researchers and artists will use or adapt his composition to support science outreach, and has released the score and sound files under a Creative Commons license.

Musicians are free to do with this what they will! And if you’ve watched Vi Hart’s latest, you know that a jumble of random notes can sound awful interesting when you put the right chords behind it. So what are we waiting for? Let’s build on this, musicians. Let’s make it pluck our heartstrings, fill us with dread, make us fear for the future. Let’s make data sonification sound like music, for once!

(PS: I haven’t actually found said score or sound files anywhere on the internet, but it’s just over 100 notes. It can’t be that hard to transcribe, people.)

Update: Commenter dodobird tracked ’em down! Here’s the mp3, and here’s the sheet music in PDF form. If you end up building on the melody, tell us in the comments!

Here’s the video: