YOU can find extrasolar planets

By Phil Plait | December 17, 2010 10:30 am

Ever dreamed of making an astronomical discovery, of finding a new planet orbiting some distant sun trillions of kilometers away?

planethuntersWell, now’s your chance. The good folks at Zooniverse have started a new project where citizen scientists can actually look for extrasolar worlds: Planet Hunters.

The idea is pretty cool: they have data from NASA’s orbiting Kepler Observatory, a telescope that is staring at one part of the sky and observing a hundred thousand stars all day and night. If a planet orbits a star — and its orbit lines up just right — the planet will pass directly in front of the star, blocking its light for a few hours. That dip in light can be detected, revealing the presence of the planet!

With Planet Hunters, you’re presented with the data, and you look for those dips. Astronomers have programmed computers to look too, but the Zooniverse people think that humans can do better. It’s not a bad assumption; it’s the devil’s own work teaching a computer to sift through this kind of data. And there are lots and lots of people out there who can do this… and I’m guessing more than a few BABloggees are in that set.

They have information on the science behind all this, and simple instructions to get you started. I played with it a bit and it’s not too hard. And it’s not at all difficult for me to think that somewhere in all that data are planets waiting to be discovered, planets the computers will miss, but that can be detected by the 1500 grams of pink computer behind your eyes.

CATEGORIZED UNDER: Astronomy, Cool stuff, Science
MORE ABOUT: exoplanets, Zooniverse

Comments (45)

Links to this Post

  1. Help Find Exoplanets on your Own! | Astrotopia | December 19, 2010
  1. sHx

    Sign me up, Scotty!

  2. This takes a bit of practice. But I’m on it!

  3. david

    I’ll be trying this when I get home. I guess there is no chance we would be able to name any planets we find? If they ever wanted to increase the man hours like crazy that would be an idea.

  4. Unaspammer

    It seemed easy enough until I got to this star. Those transits are completely obscured by noise. How the heck is an amateur supposed to spot them?

  5. Mike
  6. Yojimbo

    Unfortunately, it won’t work with my browser.

  7. Mark T.

    I found the opening page a little bit of a disappointment. Why would you go live with a site that does not work with Internet Explorer? The comment about working with “a modern browser” might be cheap.

    Wait, I just double-checked the site and it now reads:

    “We’re sorry – our site doesn’t yet work in Internet Explorer. We hope to fix this early in 2011, but for now download one of these modern web browsers instead”

    A little better, but still…

  8. A Mormon is rewarded after death with jursidiction over a planet. This suggests an untapped, well-organized and -funded, intelligent target audience eager to re-establish ties with ancestors.

  9. Beau

    This is totally awesome. I love everything that Kepler is doing right now. In fact, I’m even naming my first born son Kepler (due in March, btw).

  10. Chris

    Darn it Phil. You crashed the site. I could see it this morning when I checked my email now it won’t load.

  11. QuietDesperation

    A Mormon is rewarded after death with jurisdiction over a planet.

    Wow. Really? Dang. Add Velociraptor Jesus to that and I might have found a religion I can love.

    Anyway, a childhood injury has allowed me to find extrasolar planets with my trick knee for years. Oh, and predict the weather.

  12. Mark McAndrew

    They need more examples of real transits. Pretty sure I’m seeing too many…

  13. Joseph G

    @#s 6 and 7: Bwaaaahahahahaha!
    I remember the bad ol’ days when some sites would ONLY work with IE. Heaven forbid anyone want to use browsers that aren’t riddled with security holes!
    Now, to be fair, IE has improved a lot over the years. I just think it’s funny that the ubiquitous browser that Microsoft got sued over (remember the anti-trust hubbub, asserting that Microsoft had a monopoly, in large part because of their inclusion of IE in Win 95 onward?) is now apparently not “modern.”

  14. Yojimbo

    They recommend Firefox and Safari among other browsers. I have actually tried using Safari on one of my machines. It mostly worked fine, but most days at some point it would start a flurry of disk activity so heavy that the computer ground to a halt for up to ten minutes before going back to normal. And I used Firefox for a while, too. I recall not liking it much, but I don’t recall why anymore. It may be better now. I guess I should go look at other options again… but you know what they say about the devil you know.

  15. Joseph G

    @#11 QuietDesperation: Lulz! Please, pleeeease tell me that that’s a real coloring book that I can buy!!

    That was my thought too, though! Heaven schmeaven, I want my own frickin’ planet! It’d be like the universe’s most awesome game of SimCity! Or Black and White,” depending on how the mood strikes you ;) I wonder if you can pick which planet you get? Can they be populated? Can YOU populate them? Can you choose who gets to, ahem, help you populate them? ;)

  16. Joseph G

    @#14 Jimbo: Personally, Chromium with minimal addons has worked best for me. I’ve noticed an odd trade-off with Chrome and Firefox – Firefox seems to use much more memory, but Chrome seems to use more CPU even when idle. I think a lot of it is due to Flash, though. Chrome comes with Flash integrated into it, which is a plus or a minus, depending on your own opinions. Personally I loathe Flash and can’t wait for it to be replaced by HTML5 and other technologies, but at the moment it’s got a frickin’ stranglehold on the market :(
    *stops and reads post*
    Wow, I’m a geek :P

  17. Francis O'Donovan

    Brings back memory of staring at computer screens during grad school. Still, TrES-2, -3 and -4 came out it!

  18. SkyDancer

    Yeah, I thought it was easy too, until they snuck in some fabricated data and I completely missed all three transits I was supposed to see, because they were buried in the noise. Their tutorial example was baby simple compared to that one.

    I was so demoralized and fearful that I’d screwed up on my interpretations of the data that I gave up forever for at least an hour.

    There’s a scientist for ya. Though I’m afraid I may be overanalyzing now…

  19. Pan

    Have you ever wondered what the “sky” would look like if you were sitting on a mu meson?

  20. Peter

    I can second the sentiments posted here.

    The training could be better. What to look for. What *not* to look for.. The site could also make more clear what’s training and what’s for real. At least, that was unclear to me.

    The first one was easy. “Let’s look at some real data” the site then said. It confused me no end. Was this now for real or still training? I saw a graph clearly showing an osccilaition in britghness. How should I interpret that? I’m just a layman. Was I looking at a binary star? A Cepheid or some other type of variable star I know nothing about? Or a big planet orbiting its star? Not knowing whether this was still training or for real I was reluctant to go forward. The third looked like noise: I could not decide either way. Afraid to pollute the database, I stopped.

    I hope they improve it because I would be more than willing to be part of a forefront of science. That’s a kinda cool, methinks.

  21. andy

    In Soviet Russia, extrasolar planets find YOU.

  22. Wayne

    For those that are worried about “polluting the database,” my guess is that they have multiple people checking each and every one of these. I looked at data for about 90 minutes trying to figure out if I was doing things correctly, when it showed me this one – I marked the dips, excited to have found a planet and clicked on Discuss, only to find that others had already done so. So, it looks like there is a lot of redundant checking going on, even for the real stuff, so I wouldn’t worry too much about missing something. Someone else out there will probably spot it.

  23. Tim G

    This is a great idea but I’m getting a lot of lag in the interface :-(

  24. Reverend J

    Toyed with it for a few hours, found several eclipsing binaries and a potential planet with a period of about 5 days, pretty cool stuff, but it does get tedious at times.

  25. Skrim

    Heh. Just began this after seeing it here, and second star after finishing the tutorial I got this:
    Hmm. Looks like an exoplanet. Beginner’s luck, maybe.

  26. Skrim (25) – that drop looks real, but may not be an exoplanet. It could be an eclipsing binary or some other effect. But I hope they follow up with it!

  27. Derek

    I’ve found all kinds of nifty stuff, including 4 or so that could be planets (out of 800+ I’ve reviewed =p):

  28. Joseph G

    Warning: major math n00b here with no programming expertise:
    Question – I know that human brains are good at spotting patterns, and that it’s difficult designing algorithms that reliably pull patterns out of noise like this, but would it be difficult to design an algorithm that sort of directs your attention to possible “hits”? Presumably, any transit is going to be reliably periodic – would it be difficult to design an algorithm that “shades” or “inverse highlights” bits of the data that don’t appear to have any periodicity?

    Also, what exactly is the source of all the noise? Does the interstellar medium produce scintillations (like the atmosphere, but presumably much slower and smaller in magnitude)? Is it cosmic rays hitting the detector? If it’s something like the latter, it seems we could make things tremendously easier if we had two or three identical scopes (yes, I know we aren’t exactly swimming in government money right now, but a guy can dream).

  29. Mike Saunders

    @28 Joseph G
    Re: Periodicity
    The Kepler measures each star for only 30 days apparently, so to see any periodicity the planet will have to be orbiting faster than that… probably many planets will be missed the first go round, and the time measured for each star will increase the longer kepler is up there, but I don’t know for sure. In our own solar system Mercury orbits every 88 days, so it would probably be possible to miss every single planet here using this data. Most of the highly periodic graphs on there are due to eclipsing binarys.
    Also the planet might transit the star for a very short time, only one or two data points showing the effect…

    Re: Noise
    Any detector has to deal with thermal noise in the electrical components. For terrestrial applications its not such a big deal for most things because other sources of noise will bite you first (Weather radar for example, terrain, bugs in the bug jet stream, birds etc) but in space it is easily the number one contributor. Cosmic rays hitting the detector do happen, but the effects are a lot bigger than normal distribution noise (they have some examples of it on the website, usually in the form of a discontinuity)
    There are many ways to reduce thermal noise, the more aggressive approaches have huge cost and complexity issues. I have absolutely no idea about the mitigation techniques used on Kepler, but considering they are measuring extremely small changes in light, it was probably an issue scrutinized closely. I don’t have any doubt that noise you are looking at is minimized to the fullest extent current technology allows (within mission parameters of course)

  30. Mike Saunders

    Re: Thermal Noise

    One way to reduce it is using very cold liquids around your components, like the Large Hadron Collider uses. But the trade off: you depend on this liquid to reduce to noise enough to get good data, if it vents into space for any reason you are screwed. Liquid cooling system are really complex. At least one mission uses liquid cooling, and it seemed to me, the nervousness during the cooling process was really apparent.

  31. Joseph G

    @ Mike Saunders: Ahh, I see. Thanks for the explanation.
    And regarding the thermal noise (I didn’t even think of that), I was just saying that regardless of how clean they’ve been able to make the data, if you had two or three telescopes to work with, you could filter out most of even THAT noise simply by overlaying the data from all of them :) Unless there’s some sort of quantum/information theory minimum noise limit that I’m missing.
    Perhaps one of these days humanity will get its priorities straight and spend as much on science as we do on bombs and various wars on drugs.

  32. Joseph G

    @# 30: I’m super-duper impressed by the amazing passive cooling/sunshade system on the JWST, but then that looks to be a really monumental undertaking, and not something you tack on to an already pricey telescope if you don’t NEED it :)

  33. Joseph G

    I’m trying it out now. This is fascinating. The big question I have is, how do you tell a regular variable star from a transiting binary? They’re both going to have very regular fluctuations.

  34. Joseph G

    It seems like many dwarfs are quite noisy. I dunno if this is simply because they’re dimmer so the signal/noise ratio is lower, or if lower mass stars tend to flare and vary a lot, or both?
    Meh. I love this idea, but I feel like I have very little chance of picking out anything real, if I saw it.

  35. Messier Tidy Upper

    Great idea. I love it. :-)

    Off topic but thought I’d share this interesting news via the LRO :

    Lunar farside in false colour, a different view of Earth’s natural satellite.

  36. Messier Tidy Upper

    @ 34. Joseph G : Yes, lower mass stars especially red dwarfs are famous for often having massive stellar flares and being variable – see :



    for more info. – & a great piece of space art in the BA blog one on EV Lacertae. :-)

  37. 24601

    It’s a cool project, it’s just sometimes astonishingly difficult to identify the transits against the variability of the star.

  38. réalta fuar

    Cloud sourcing has worked remarkably well for seti@home and several other projects, it’s going to work remarkably poorly here. Humans don’t do EVERYTHING better than good software, and this is one of the problems where computers are MUCH better (though I’m sure all transits from all programs are checked by eye and brain). Look at some of the WASP transits for lots of good examples where a human would have little or no chance of finding the real transits. Not to mention the fact that a) the Kepler software is very, very good as they’ve had years to get it right and are constantly tweaking it and b) humans are going to find lots and lots of false positives, even more than the software finds. Checking false postives wastes incredible amounts of resources.
    c) the Kepler team has been extremely reticent about handing over good data, does anyone seriously think they’d give out data (for transits) that hasn’t already been thoroughly picked over? It’s nice to get people excited about science, but the chances are this isn’t going to work out well in the long run.

  39. lame, site doesn’t work in IE9 and tells me to get a modern browser lol

  40. fred johnson

    Internet explorer is a piece of crap that’s why , and it’s not a matter of opinion….

  41. Mick

    @39: The main purpose of the site isn’t finding planets (where computers are better at), but categorizing different star types.

  42. Will I be able to name a planet that I find? Is that how the naming process works?

  43. Joseph G

    @#39 realta fuar: I have no expertise in this area, but my gut feeling tells me that you’re right. Isn’t picking a coherent signal out of noise pretty much exactly what the SETI@HOME project does? Wouldn’t it make just as much sense, instead of the distributed processing that the project does, to have people view (or even listen to a frequency corrected version of) the SETI@HOME data?

  44. Joseph G

    @#43: I seriously doubt it, unfortunately. For a “hit” to come to the attention of the researchers, I’m sure it has to be flagged by a number of different users, so right there you’d have a problem.
    And even when scientists do discover exoplanets independently, they still have pretty boring names (typically “[star name] [lowercase letter of the alphabet starting with B based on number of objects found] e.g. “HR 8799b”


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!


See More

Collapse bottom bar

Login to your Account

E-mail address:
Remember me
Forgot your password?
No problem. Click here to have it e-mailed to you.

Not Registered Yet?

Register now for FREE. Registration only takes a few minutes to complete. Register now »