Data to Desktop
Aug 9th 2007Wallabyeyes & nerd alert
One of the more interesting things to do with data – at least certain types of data – is to look at it. As the “Small World” Photomicrography competition puts it, “…a good photomicrograph is also an image whose structure, color, composition, and content is an object of beauty, open to several levels of comprehension and appreciation.” Some kinds of data don’t really lend themselves to being treated this way, of course – a straight line is a straight line.
Other datasets take a lot of work to turn into something that can be “looked at” at all. This was much more of a challenge with the Raman data I used to collect: 60,000 noisy spectra aren’t really all that attractive. By looking at the parts of the spectra that changed based on the sample – for instance, a blob of plastic embedded in glass – it’s possible to convert this big pile of data into a “chemical image”: A two-dimensional representation of the chemical features of the specimen.
My current data tend to be much more in the “straight line” category, unfortunately, so there aren’t as many opportunities to turn the data into pretty pictures, or even interesting ones. Sometimes it’s easy to play around a bit and come up with something after all. I finally got the interferometer set up again, and used it to produce this picture…
An interferometer is a device that… well… measures interference. Light travels as a wave, of course, and one property of a wave is that it’s got peaks and valleys. If you take a wave, and a copy of the same wave, and stack them up so that the peaks align with each other, then the peaks get higher and the valleys get deeper. This is called “constructive” interference – we’re building the wave up. If you shift the copy of the wave over, so that the peaks are on top of valleys, then you’ve effectively “filled in” the wave so that it’s a flat surface again. This is “destructive” interference.
The interferometer is a device that splits a light wave in half, and lines up the two copies at different offsets. Then we measure the intensity of the combined waves, and plot the intensity as a function of the offset (the position of the copy). Why do we want to do this? Among other things, it lets us measure the length of a very short pulse of light – that was why I was setting it up this past week.
That’s what a single pulse looks like – the whole thing lasts only about 30 femtoseconds. (1 femtosecond is one-millionth of a nanosecond.) That’s the time it takes light to travel a distance equal to about 80 human hairs.
So how to go from that pulse to the picture? Simply blow it up to two dimensions (by stretching it “out of” the page), then flip a copy of that around, tint one copy green, the other blue, and combine them. So this is the “interference” of the interference pulse… interference squared?
Currently my desktop background… got larger copies too, but you’ve got to let me know you want one. Comment, people! 