Pretty though data can be, there's no subsitute for a physical model sitting on your shelf. Particularly if that model is a three kilogram glass brick that, if neccessary, can be used as a blunt-force instrument against anyone trying to steal your precious data. This is what's been termed a dataweight :

Inside the glass (14x9x9 cm) are about a hundred little cigar-shaped blobs. These are hydrogen detections of galaxies in the Virgo Cluster. The reason they look elongated is because the telescope doesn't have enough resoluton to map structures within each galaxy. So in a normal, 2D map they'd just look like points. Fortunately we also get a third axis with this data, which is not distance, but velocity.

On the very large scale, velocity turns out to be a pretty good proxy for distance (though it gets a bit messy). But on the scale of an individual galaxy, it doesn't really relate to distance at all, since galaxies are rotating. While the telescope doesn't have great spatial resolution, it does have fantastic velocity resolution. That means each galaxy may only span a few pixes across the sky, but it velocity it can span tens of velocity channels. Hence the appearance of blobs.

This particular data set is the heart of my thesis (actually it's about a quarter of all the data for my thesis, but this is by far the most important bit). To create the cube, first I converted the radio data FITS file into a simple text file containing the x,y,z positions (where x,y, are Right Ascension, Declination positions on the sky) and S/N (signal-to-noise, basically, how bright the detection is). You can read more about that here.

There isn't really any way to etch the brightness of a particular detection into a glass cube. The process involves focusing a laser at many thousands (or tens or hundreds of thousands) of points in the glass, creating microscopic fractures. You either create a fracture, or you don't. So from my text file, I removed all the faintest features (noise) leaving only the galaxies. And then I sent the file in the direction of Bathsheba Sculpture.

Some time later my dataweight arrived in an elegant, velvet-lined box. An extensive bout of google searching eventually found a perfectly-sized light base. And that really shows things off properly. Comes complete with disco mode !


The cube looks beter from one direction than the other because the fractures aren't symmetrical, so they reflect light differently in different directions.

The cube normally lives on my desk, where the backdrop of my thesis entitled "The Virgo Cluster Through the AGES" forms a perfect label. It makes a great teaching aid and presentation showpiece. The only problem is that unlike regular data, it's difficult to transport. You can put it on a memory stick, but all this will do is crush the stick. And I dread to think what airport security will say if I try to take it in my hand luggage.

rhysy.net | feedback@rhysy.net