Reinventing Discovery - Michael Nielsen [69]
Galaxy Zoo Revisited
I can honestly say that Galaxy Zoo is the best thing I’ve ever done. . . . I don’t know quite what it is, but Galaxy Zoo does something to people. The contributions, both creative and academic, that people have made to the forum are as stunning as the sight of any spiral, and never fail to move me.
—Alice Sheppard, volunteer Galaxy Zoo moderator
Galaxy Zoo began in 2007, with two scientists at Oxford University, Kevin Schawinski and Chris Lintott. As part of his PhD work, Schawinski was looking at photos of galaxies. Galaxies come in many shapes and sizes, but most galaxies are either spiral galaxies, like our own Milky Way, or else elliptical galaxies, roughly spherical balls of stars and gas. Conventional wisdom in 2007 held that most of the stars in elliptical galaxies are very old stars, getting up toward 10 billion years in age. When stars get old, they will often change color and size, turn into red giants, with the result that many elliptical galaxies have a reddish tinge when compared with spiral galaxies, which are younger, and contain many newly formed blue stars.
Schawinski suspected that the conventional wisdom was wrong, that some elliptical galaxies might not be so old after all, and there might be a lot of star formation going on inside them. To test his suspicion, Schawinski spent a week poring over photos of 50,000 galaxies from the Sloan Digital Sky Survey (SDSS), looking to see which of the galaxies were elliptical and which were spiral. As I mentioned in the opening chapter, distinguishing elliptical and spiral galaxies is something humans still do better than computers. Once he finished the classification Schawinski used a computer program to analyze each elliptical galaxy, to see how red or blue it was. As he had suspected, the results suggested that the conventional wisdom was wrong, that star formation was going on in some ellipticals. Unfortunately, the effect was weak, and he needed to analyze a much larger sample of galaxies to really nail it down. Fortunately, as we discussed in the last chapter, the SDSS had made images of 930,000 galaxies openly available. This was a promising but daunting resource. Classifying the first 50,000 galaxies had involved a heroic weeklong effort by Schawinski—to classify 50,000 galaxies over seven 12-hour working days requires classifying an image every six seconds! Even at that tremendous pace, it would take many months to classify 930,000 galaxies. And there’s no way Schawinski could maintain that pace. Even if he devoted most of his working time to the classification, it would take years.
One day in March of 2007, Schawinski adjourned to the Royal Oak, a pub in Oxford, together with a postdoctoral scientist who had recently arrived at Oxford, Chris Lintott. Over a pint they considered a wild idea for classifying the SDSS photos. Instead of doing the classification work themselves, perhaps they could build a website that would invite members of the general public to help out. They dragooned some friends who worked as web developers to help build the site, and on July 11, 2007, the Galaxy Zoo site went live with an announcement on BBC Radio 4’s Today program.
The response to the announcement of Galaxy Zoo dwarfed expections, overwhelming