The return of citizen science
The internet provides the means for the keen amateur to once again contribute to scientific inquiry, writes Darren Baguley.
It was Saturn’s golden glow filling the view of his friend’s telescope that awoke Trevor Barry’s passion for the night sky. So one morning, when the retired miner reviewed images of the planet he had shot the night before, he knew straight away there was something different: “There was a very small white spot and I knew it shouldn’t be there.”
It was February 23, 2008, and his hobby was about to change his life. Checking his reference books, Barry realised he had captured a rare event: one of a handful of Saturn storms over the past 200 years powerful enough to be observable from Earth. The white dot was a weather event thousands of kilometres in diameter, with lightning 10,000 times more powerful than anything ever witnessed on our own planet. Had any of the professionals spotted it? Barry found nothing on the NASA Cassini space probe website, even though the craft was orbiting Saturn and its moons. So he “pestered people”.
Having completed an online master’s degree in astronomy from Swinburne University of Technology in Melbourne, Barry turned to his academic contacts. One put him in touch with the Cassini mission imaging team. Looking back through the data, the scientists noticed Cassini’s radio wave sensor had recorded a telltale blip. Barry was right.
Impressed with the quality of Barry’s images, they invited him to become the fourth member of a select group of amateur astronomers supplying the mission imaging team with data.
“I’m the most prolific collector of this data on the planet because of where I live and the number of clear nights I get,” says Barry, who lives in Broken Hill in far-western New South Wales.
In the 21st century, anyone with a laptop and an internet connection
can join the growing ranks of the citizen scientists
Articles using his data have since appeared in journals such as Nature and Science. He is now as busy mining astronomical data as he ever was when he mined zinc from the Broken Hill ore body. Barry has become a poster boy for citizen science, a throwback to the gentleman scientists of the 19th century and earlier. Back then eager amateurs made major contributions to science. They collected fossils and discovered dinosaurs, observed the night sky and named new comets, invented a microscope and discovered bacteria swimming in a glass of water. All they needed was the time and inclination.
But pushing back the frontiers of science in the 20th century required immensely sophisticated and expensive instruments – carbon-dating techniques to study fossils, radio telescopes to search the universe, particle accelerators to probe matter. Amateurs were locked out.
The 21st century, however, has put amateurs back in the loop. Anyone with an internet connection can join the growing ranks of the citizen scientist. Not only that, the scientists need them desperately. Every corner of the universe is being explored by telescope or monitored by satellites, environment sensors or wildlife video cams. Scientists are deluged with data. Many are coming to rely on passionate volunteers willing to lend a hand.
Nearly a decade ago I joined the search for extra-terrestrials. SETI@home, a citizen science forerunner still functioning today, simply involved downloading a program that enabled a home computer to use its spare capacity to process data from the Arecibo radio telescope in Puerto Rico, which listens for radio transmissions that might just be coming from ET. The project hooked me from the start. The SETI screen saver was hypnotically beautiful, a colourfully writhing real-time graph of my computer’s data-crunching exploits. With the hope it might be my PC that found ET, I bought ever more powerful computers. I signed up to other “distributed computing” projects such as ClimatePrediction.net, which aims to test and improve climate models. I tell my wife my overclocked dual-graphics-card monster computer is for playing games; the truth is I am addicted to citizen science.
Science’s success at co-opting people’s computers inspired the next stage: co-opting people’s brains. First cab off this rank was Stardust@home. In 2004, NASA’s Stardust spacecraft intercepted the comet Wild-2 and captured thousands of comet dust grains using a collector filled with a state-of-the art aerogel that captures particles whizzing by at several miles per second. Among these were flecks of something even rarer: 100 specks of stardust. Formed in exploding stars, this is where most of the heavy elements reside in interstellar space. It is this stardust that formed our solar system and everything in it. To study it, though, required sorting it from the comet dust captured in the aerogel – a task akin to searching for ants in a football field, five square centimeters at a time. That’s where citizen scientists lent a hand on a job that would have taken the researchers decades to do on their own.
But the project that really captured the public imagination was Galaxy Zoo, an effort to classify more than 930,000 galaxies captured in three-dimensional maps by the Sloan Digital Sky Survey (SDSS) over eight years. Classifying them by shape was considered a good start, since a galaxy’s shape tells its story. As they age, galaxies change from fresh young spirals actively forming new stars to mature spherical “ellipticals”. Odd bumps and clumps can indicate collisions and mergers. Computers, however, floundered at this kind of pattern recognition task.
Contemplating the scale and mind-numbing dimensions of the job in a local pub after manually getting through 50,000 galaxy classifications – just 0.5% of the available data – University of Oxford astrophysicists Kevin Schawinski and Chris Lintott struck upon the idea of emulating Stardust@home.
“It was sheer desperation, really,” Lintott recalls. “I thought, if people are prepared to look at dust, surely they’d be prepared to look at galaxies.”
Galaxy Zoo’s premise was simple. When someone registered online, they viewed their first SDSS image and answered a series of questions. Was it a smooth circle, oval or cigar shape? A disk or spiral? Clumpy or irregular? Perfectionists could agonise over images, although according to Lintott: “Usually your first impression is right.”
‘If you can engage the public then they come to your support in times of need'
Within weeks of starting, the project had garnered 80,000 volunteers who classified more than 10 million images, with each image classified by multiple volunteers. The project was soon complete. Then came Galaxy Zoo 2, where volunteers helped with more detailed examination of 250,000 of the brightest galaxies from the original project. That phase was declared complete after collecting 60 million classifications. Galaxy Zoo is now well into its fourth phase, with volunteers processing the million-plus images from the Hubble Telescope CANDELS (Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey) program, studying galaxies so far away that the light we see was emitted when the universe was less than 4 billion years old.
Citizen scientists are motivated by lofty notions. Not only are we learning science in leaps and bounds, in the back of one’s mind there is also the possibility of discovering something really remarkable. Like Hanny van Arkel, a Dutch schoolteacher who became involved with Galaxy Zoo after reading about it on the website of astrophysicist and Queen lead guitarist Brian May. In 2007, she reported a strange object (or “voorwerp”) glowing green on her screen. Nothing like it had been seen before. It quickly became known as “Hanny’ s Voorwerp”.
Lintott suggested it might be a gas cloud lit by a nearby quasar, a fantastically bright ball of hot gas with as much energy as hundreds of galaxies, surrounding a supermassive black hole. But no one knew for sure. The Galaxy Zoo team got some telescope time on the Westerbork Synthesis Radio Telescope in the Netherlands to study it more closely. The pattern of signals supported Lintott’s interpretation. A quasar’s bright emissions were lighting up the gas cloud, causing oxygen within it to glow bright green.
Hanny’s Voorwerp has featured in some beautiful astrophotography, and astronomers have gained useful new data from it. It seems to overturn theories that quasars slowly shut down over 100 million year timeframes.
While the Galaxy Zoo project has put citizen science on the map, not all astronomers are convinced about the project’s value to science. “The science benefit is arguable, I’d say,” says University of Sydney astronomer Joss Bland-Hawthorn. He says the latest software could probably do just as well as people can.
The benefit of software is that it lacks the biases of human data processors, says palaeontologist Paul Willis. Now director of the Royal Institute of Australia, Willis recalls looking for fossil teeth in the Riversleigh deposit in northwest Queensland. “We were getting vast amounts of sediment to be sifted through, mainly looking for small mammal teeth.” To help, volunteers were trained and let loose. But many missed important fossils or, just as annoyingly, wasted the scientist’s time with irrelevant findings.
“When you’re dealing with fossils it’s relatively easy to pick those biases,” Willis says, “but when you’re talking about people sifting through hundreds of thousands of photographic images, I could see the biases of identifying junk or not identifying important stuff being rampant and very difficult to control for.”
Lintott says Galaxy Zoo took such issues into account right from the start, with accuracy being checked by having up to 30 people looking at each image. He also disagrees that it is impossible to control for bias. “We can measure bias. I know that people miss 27.43 per cent of the spirals. We take it very, very seriously.” Crowd sourcing is not a panacea, and there are still tasks that can only be done by professionals: “But if machines can do 70 per cent, and the crowd can take most of what’s left, you can do some very good work,” Lintott says. And it’s not just a matter of getting the science done. “The public outreach side has been an inspiration, and that’s really crucial,” says Bland-Hawthorn. “If you can engage the public then they come to your support in times of need.”
Bland-Hawthorn says biologists or chemists will often say their first love in science was astronomy. “It brings people into the scientific world … Outreach is a really important role and Galaxy Zoo has really nailed that,” he says. Since Galaxy Zoo’s success, crowd-sourced projects have mushroomed, many hosted by Lintott’s team under an umbrella organisation called Zooniverse.org. Although the new projects stray far from astronomy, Lintott is happy to help out with his hard-won expertise. “It’s not just a case of putting the data up on a website and reading the answers off,” he says.
In many ways, the rise of the citizen scientist has brought science full circle. Early 19th century German astronomer Friedrich Bessel pre-empted the Galaxy Zoo project by 200 years, says Gerhard Wiesenfelt, who researches the history and philosophy of science at the University of Melbourne. The first high-quality commercial telescopes had just appeared, and amateurs were snapping them up. Bessel tapped this resource, giving like-minded amateurs a quadrant of the sky to chart, then stitching together their work to create the first high-quality star map. “It was the kind of thing that was quite happily done in those days,” says Wiesenfelt, who sees today’s citizen science projects as part of a broader movement to democratise science.
Australia’s best-known professional astronomer, Fred Watson, is enthusiastic about the process. “People like Trevor Barry are collecting data that would simply not be available otherwise,” he says. “I think we’re going to see a much greater involvement with Joe Public in the scientific research that’s going on. It’s a great trend that will take us into realms we couldn’t have entered otherwise.”