As an astronomy graduate student at Oxford in 2007, Kevin Schawinski faced a rather sizable problem: He needed to sort through 50,000 images of galaxies taken by a robotic telescope one by one, and classify each galaxy according to its shape.
After spending a week of doing nothing else, Schawinski was sure there had to a better way.
Sitting in a pub one night, he complained to colleague Chris Lintott, a postdoctoral researcher at Oxford, about how long it would take him to go through even a fraction of the images. It became clear he needed help. So, with the aid of Lintott and several others, he created Galaxy Zoo, an online citizen science project that asks anyone interested to donate some of their time to look at galaxy images and help classify them.
"I thought there were maybe a couple of dozen hardcore amateur astronomers out there who might possibly be interested in this," says Schawinski, now a postdoctoral associate in Yale's Department of Astronomy.
He was wrong.
Thanks in part to a story by the BBC, where Lintott co-hosts an astronomy program, thousands of users were busily classifying tens of thousands of galaxy images within the first 24 hours of the site's launch in the summer of 2007. Caught totally off guard, the overload on the server storing the images actually melted a cable and nearly threatened the whole project. "We were overwhelmed," Schawinski says. "It was a complete surprise."
A year and a half later, the numbers are even more overwhelming. Galaxy Zoo is thriving, with more than 180,000 users who have made nearly 75 million classifications of one million different images — far beyond Schawinski's original 50,000. If Schawinski was still struggling on his own, it would take him 124 years to classify that many images — proving, he notes, what a powerful collaborative tool the Internet can be.
"We never thought of it as an outreach project, we never thought of it as an education project. We always thought of it as, ‘Please help us with our research, in a way that only you can.' And that really attracted people," he says.
A project like this one requires using actual people instead of a computer algorithm, Schawinski explains, because computers aren't very good at pattern recognition. Humans, on the other hand, are exceptionally good at picking out different patterns and shapes, such as being able to distinguish between people's faces at a quick glance, he notes. The same turns out to be true when it comes to distinguishing different types of galaxies.
The premise of Galaxy Zoo is simple. Users are shown an image of a galaxy and asked two basic questions: Is the galaxy an elliptical galaxy (a type of galaxy with no dust or gas, but many stars) or a spiral galaxy (with rotating arms, like our own Milky Way galaxy); and, if it's a spiral, in which direction are the arms rotating? In order to participate, users take a 10-minute tutorial teaching them the basics of galaxy morphology.
At first, Schawinski was skeptical as to how well the average person would be able to classify the galaxies. In fact, he says, members of the general public are often better than professional astronomers because they don't have the same background and preconceptions about what they're seeing.
He was also surprised at just how many people are interested in donating time to the project, noting that users range from individuals with no astronomy background, to school teachers and students, to parents who participate with their children as a sort of family activity.
But Galaxy Zoo is about more than just looking at pretty pictures of galaxies. The project has resulted in real scientific discoveries, with several papers already published using the data and a dozen or so more on the way. The Galaxy Zoo team - which includes astronomers from Yale and Johns Hopkins University in the United States, and the University of Oxford and the University of Portsmouth in the United Kingdom - has often been surprised by the results.
Bill Keel, an astronomy professor at the University of Alabama who studies overlapping galaxies, decided to ask Galaxy Zoo users to contact him if they came across an example of this rare phenomenon. Throughout his career, Keel had studied the dozen or so overlapping galaxies then known to astronomers. Within a day of posting his question on the Galaxy Zoo forum, he had more than 100 responses from users who had indeed found such objects. Today, thousands have been identified.
"It's a powerful multiplier," Schawinski says. "It makes research possible that just wasn't possible before."
Another interesting object found by a now famous Galaxy Zoo user is Hanny's Voorwerp - a cloud of dust still being illuminated by the light from a nearby quasar that has since died out. The object was the first of its kind; it was discovered not by a professional astronomer, but Dutch schoolteacher Hanny van Arkel.
"Computers cannot recognize the unusual. Humans can," Schawinski says. "Hanny's Voorwerp would never have been found except for a human looking at it and saying, ‘That looks really odd and unlike anything else I've seen before,' and alerting the community to it. And it turned out to be an incredibly exciting object."
Perhaps one of the biggest surprises, however, is that the science coming out of the project sometimes goes beyond the world of astronomy altogether. One of the reasons Galaxy Zoo asks users to assess the direction of rotation of the spiral galaxies is because astronomers have discovered more spiral galaxies with arms rotating counterclockwise than clockwise - a result that would turn everything astronomers know on its head if proven true.
Galaxy Zoo showed that, in fact, there are an equal number of clockwise- and counterclockwise-rotating galaxies and there seems to be a psychological effect that makes people see the arms as rotating counterclockwise even when the opposite is true. The phenomenon has interested neuroscientists from around the world, several of whom are now conducting experiments to try to explain the effect.
Now the team plans to take the project even further, with the launch of Galaxy
Zoo 2. The second generation will ask users much more detailed questions about a quarter million of the brightest objects in the image bank.
"Galaxy Zoo 1 was very broad. We realized there are a lot of questions we could have asked but didn't," Schawinski says. "That's where Galaxy Zoo 2 comes in. It allows for all sorts of new questions to be asked."
Some of those questions involve whether the galaxies appear to be edge-on, the shape of their central bulges and how many arms the spiral galaxies have — just some of the observations that will allow the team to better understand how galaxies form, merge and evolve. "Galaxy Zoo 2 will open whole new avenues of research," he says.
Schawinski says users were clamoring for a new version that would allow them to make more detailed analyses and are eagerly awaiting the chance to contribute to new science.
"They're our co-investigators — they're scientists, too," Schawinski says. "And they really care about the science they've contributed to."
To find out more about Galaxy Zoo 2 and how to participate, visit www.galaxyzoo.org.
— By Suzanne Taylor Muzzin