 It's fantastic to see such a large audience here. When I saw the Astronomy Mini Conference being advertised, I'm very delighted to be accepted to give a talk, but I wasn't entirely sure how many people from the Linux community would want to sit through what is predominantly a science session. We're going to do our best as scientists to, well, at least I will and my students will, everybody else's, as well try and draw what we do into the realm of the conference being Linux and FOS as best we can. So please forgive me if I sort of drift into the science realm. Stay tuned, I will bring it back to the topic as best I can. I give this my title to this talk, Visualizing the Open Universe, a somewhat grandiose title which is perhaps a little bit misguided. However, the main point is, which I wanted you to take away from the talk is, as you'll see throughout the talks in this session, there's a lot of data out there coming from the astronomy groups, both present and planned. We're basically being firehosed with data. And the common refrain I hear from colleagues in my own group and other groups is, we don't have enough people slash funding to analyze all the data coming towards us. So this is a problem, it's a problem now, and it's only going to get worse. When the SKA comes online, when things like the LSST, this big synoptic survey telescope comes online, there's going to be even more data, the problem is going to be compounded. So part of the work which I'm interested in doing is seeing how you can package these data to present to people who may not necessarily be experts in the science. You've seen the sort of thing before in terms of science, Chris Linton, Galaxy Zoo, that sort of thing, hugely successful. How do you take data which we are generating and give it to folk who may not necessarily know their coordinate system from the exhaust pipe manifold or whatever, doesn't matter. We need them eyes on a computer screen to make human decisions. And if there's one thing which I've learned, I'm sure you guys know, it's extremely difficult to make a computer behave like a human. Give it some data, noisy, gappy, nonsense data and say, is this a picture of a cup or a kitten? Depends, right? You give it some scrappy, some dust on my primary mirror. Difficult, right? Getting a human to make those sorts of decisions. Well, it's easy if you've got a lot of humans. So the primary goal of my talk is just to get an idea of where I'm standing. I'm no expert in the field of open data, certainly. I'm not a database expert. I'm certainly not a computer scientist for reasons which you will see hopefully. So I'm just going to give a brief overview about me, where I come from and the project that we're doing. Talk about my particular research interests and I'm in a happy situation where I can hand over the discussion of the science to some very gifted students who are in the room who will be talking this afternoon. I get to take this overview managerial, which is fantastic. It's the first time in my life I've been able to do that. I'm usually the person who's saying, oh, this is what we did. Now I can say this is what my students are doing. That's fantastic. So a little bit about me. I did my PhD here at Auckland. I went to Dodger Bank in Manchester for a postdoc. The funding went even more south than it usually is in the power of state. Then I had to start retraining as a patent attorney. That didn't work out too well. So I skipped into back into academe. And now I'm lucky enough to get one of these rather for discovery fellowships to work back in research. My particular research interests in the past I've looked at variable stars. These are stars which pulsates, otherwise vary their output during time. I've looked at the structure of our galaxy and I've looked at high proper motion stars, stars which appear to move very, very quickly. It's interesting sort of collections of stars. What I'm presently interested in is discovering planets. It's like Pauline Arnaud in Victoria. I'm also interested in looking at gravitational waves, prediction from general theory of relativity. If you've got two really heavy things crashing into each other like two neutron stars, you'll get waves in the very fabric of space-time. So cool being astronaut being able to say waves in the fabric of space-time. Who wants to be a patent attorney? As Pauline warned you, there will be a fair amount of this particular scientific study called microlensing. This is the field which I work in. That's what I did my PhD in. It won't solely be on microlensing. I'd be pleased to know because that gets a bit dull after a while, even though the rest of us who are working in the field. However, it's an important one for you to be aware of because it's a field in which New Zealand is contributing at world leader level. So it's important science for this country. Mining, not in the Australian open-pit sense, but in the database realm, I'm interested in getting into the concept of, well, we're collecting all these data. There are discoveries to be made in those data. And those discoveries are relatively easy to make if you know how to ask. So this brings me on to a large section of the talk, perhaps, using the data that we have already to make immersion discoveries, serendipitous discoveries. Lucky discoveries, if you like. Value-added discoveries. We set up these massive projects to do very specific tasks. We get our data. We say, yes, we did, or no, we didn't. And then we have these, when we call massive, this is probably not massive to you guys, large data sets in which there's also other discoveries to be made. But writing a research proposal to a government to say, hey, we've got all this data, we want to look for cool stuff, because we know it's in there. We think, it doesn't really work that well. They want, you know, you are definitely going to discover something. Maybe. It doesn't really work that way. This is pure science, not engineering. It's different. So I'd like to just have a quick slide about my Linux credentials, such as they are, because it does mean something, I think, when you're talking about people, when they're working and computing, and any kind of endeavor that they do, where they came from, how they got to where they are. And so my particular story goes like this. My mum and dad, who were living in Australia at the time, bought me, I suspect, secondhand from a garage sale, one of those, yes, does anybody recognize that? Just out of interest. Sorry, I heard it. It's a wizard. It's a wizard, exactly. You're the only person I know of who actually recognizes this. That's all. This is, ladies and gentlemen, the Dick Smith wizard. Glorious machine. I think my parents gave it to me just to shut me up, because it played games, but I also had a little cartridge plugged into the machine, and now you start coding basic. And that was it. From ten years old, I was me coding away in basic. A porting code, actually my skills haven't really improved. It's just a language. Moving on to MS-DOS, to Windows, which I abhorred. Moving into Debian now at university, this is fantastic. And so using Debian Linux, as I do my PhD, why did I use Debian Linux? It's because, well, the person with whom I was working on the project said, I use Debian Linux, and therefore, you shall as well. Just a PhD student. Why was I so surprised? I was like, what is this Linux thing? It's like MS-DOS. It's like Command Line. That's awesome. So I just did what I was told. The work we did for a PhD was based on his, in this case, is Peter Dobbschani. He was a PhD student here working in computer science slash mathematics, but he was a brilliant computer scientist in his own right, and he developed a Beowulf cluster of Dell Optimax machines, and we used that to discover planets. Kind of cool. What was even cooler was this was a Beowulf cluster. Back in the day, you use whatever machines you had, and those machines were essentially undergraduate computer laboratory machines when the students weren't learning their stuff. So this guy managed to work out how to link these all together to a cluster with his roll your own codes, which is absolutely brilliant. It worked extremely well. He made it as part of his PhD work, he goes PhD, and then we have this cluster available for use. So I put my hand up and said, can I please use it to this planet stuff? It worked extremely well, and it was called Kalaka, which I believe is a Hungarian term for the process by which villages all get together and combine their efforts to complete a common task. So that worked extremely well. I programmed in C, but pretty much still programmed in C and owns it. Can anybody guess why that would be? That's a good answer. That's not the one I was after. Anybody else? That's what the libraries are written in. That's another good answer. Anything else? Pardon? Faster. Faster? Anybody else? I love you computer scientists. Yes? I like that answer. Easier to spell. Easier to spell? That's good. Yes? We'll take three more, because these are great answers, yes? That's the right answer. Somebody was paying attention. Debbie and the Knicks, because I was told to. You will program in C and only C, because that's what the code might do. So that's how it works. So this is an interesting question you might want to have in the back of your mind when you're training your colleagues or your students or whatever. People you come in contact with has asked them, well, how did you get into where you are? And the story I have, this is an interesting anecdote, number one, is when I was working at Jodro, I was working with a PhD candidate at the time. He was not my PhD candidate, hasten to add. He was working with a common boss. But he was a mature student, however you want to describe him, but he had worked his career as a computer scientist in the industry. So he came to us and he says, yeah, made my money, made my career. It's all good. I want to do some astronomy. So he came to us. And my boss said, you go and write up some code to model the galaxy, any galaxy. So he went, fine. And what language would you like that to be in? And my boss went, hmm, C. And the PhD candidate said, well, I don't know C, but it doesn't matter. And that's where I knew that's what a computer scientist was. It didn't matter. And he then proceeded over the next three years to write the most beautiful code I've ever seen in my life. And that's when I realized that computer science or computer coding is an art. It is an art form which can be beautiful. I have not achieved this. I will not show you my code. It is horrible. But it is commented. Thank you. So I can appreciate the abilities of a good computer scientist. And I try and impress this on people who aren't familiar with what computer science is because we get students like that. We work in physics. It's the interface, mathematics, computer work, programming. Here's some data. We use a computer. We don't use graph paper anymore. We use a computer. We write code to fit that straight line. Not a ruler and the eye. Discuss, not now. But the important thing is being able to see that computer programming is something which becomes something more than just the language that you're in. Fortunately, I've not progressed any further. I know there are such things as Python and Poole and I can even read these things. Anyway, I move on. The cluster was comprised of lab machines running after hours. As I mentioned, there's something very special about driving out to a remote campus at 7 p.m. to restart 300 machines by hand. And this is, you know, leads me on to the next point. This was off the grid. Excuse the term. This was a project that the computer science authorities, the computer science help folk knew about, but we're not prepared to, you know, this is your project, your Bay Wolf cluster, not ours. We're not going to help you. We'll make sure the machines turn on when you push the button and we'll talk about networking availability and stuff like that. But if the thing falls over, you made it fall over. These days, so yeah, unofficial, a community project, which was still a huge success. These days, we use the HPC machines at availability and the situation in 10 years has changed dramatically. I now have a name of a person and a phone number who I can call for help. Right? This is a sensational advance, highly to be commended. When I was still working, let's go back six, seven years in the UK, we hadn't quite evolved to that point. So the computer science help that was available online, web pages, written by computer scientists for other computer scientists. Yes, there are sniggers out there because I think we realize that being told of minutiae, the detail of how this thing works, I need to run some code. This is me as a user. It's not working help, please. It got to that point. It got to the point where it was easier for me to get an account, rather use my colleague's account at another institution by calling him and saying, can I please use your account because I cannot understand this web page for getting through this Byzantine login system. It's just too confusing. It was easier for me to use this account at another institution. Not good. Things have now got much, much better. Like I say, a name on a card for a computer scientist understands everything that I'm trying to do and will even offer to debug my code. I haven't tried him on this because I think the relationship will go sour. Like I mentioned, the code isn't pretty. And it works. So going from this unofficial computing project through to the work that we did for me and successive students here at Auckland, a guidance from my PhD advisor, Phil York, the science has now evolved to a degree. So that's me. That's my PhD. Evolved myself to using Ubuntu. I don't know why. I like it. There was an unfortunate incident when I had to leave astronomy and find another job. So I started training as a patent attorney. Never got there because I was about to die from distress. But that meant I had to start working with this again. They very nearly lost their computer a number of times which they gave me because it was slow and it was irritating and I just didn't have the patience for it. However, the work that I was doing for them put me into contact with strange familiar places that started to work. One of our clients was Red Hat. So the experience I had working with him, apart from a newfound interest in how the law works in particular, maybe not the practice of it so much, but how patent law works in particular was fascinating. But learning about the role which patent law works or doesn't work in a free and open source community, that was an interesting experience. So that's a little bit about me and my background going from the wizard through to Ubuntu. Let's talk about a little bit of the present work which I'm doing. And like I mentioned I'm in this happy position of being able to add talks of students who will be talking later this afternoon. The broad umbrella if you like, the data are coming from is this group, Paulie mentioned it in her talk, it's the micro-lensing observations in astrophysics group. It's a combined Japan and New Zealand collaboration. It started in the early 90s and it's still going with data here, telescope down the South Island more on it later. But the overall where is all this coming from is this collaboration, the mock collaboration. So, as an advertisement, Ashna later on will be talking about detecting extrasolar planets using this new fangal technology of GPUs and it's going extremely well. So I recommend that talk to you. Alex will be talking about discovering extrasolar planets through this particular technique of eclipse timing. Nothing to do with micro-lensing. However, we're using the same data. Again let's go back to this open data so to speak. Squeezing more science out of, not just finding what we set out to look for. Martin will be talking about automatically classifying data from these large data sets using open source software, in this case using the WECA suite of tools. So please do attend those talks and take a look at the science because all you're getting from me is a lot of some amusing anecdotes I guess. So again about the model project. This is some more advertising. Here's what we do. We look for planets. Planets are mainly discovered through these techniques. Through transit technique, the radial velocity technique, there is no test. Don't worry. The micro-lensing technique which is what we use in direct imaging. Here's what we've done. Well here's what the community, as in humankind have done so far. On the y-axis we have planetary mass. The planet's mass in units of Jupiter. And down the bottom here we have the distance of the planet orbiting from its host star. So there's Jupiter there at not surprisingly one Jupiter mass orbiting at Jupiter's distance around the Sun. And all the black dots are those planets found by various techniques by other collaborations around the world. There's Earth and we're looking for Earths. So we, as in humankind are doing quite well. If you've been paying attention to recent media, we're finding planets similar to our own Earth in terms of mass and distance away from its host star. Have we found life? No. Have we found aliens? No. Right, that's out of the way. What we're trying to do of course is trying to dig down towards this region here. I should say, have we found aliens? Have we found life? No, not yet. We're trying to carve out this space here, getting closer and closer to Earth. Just as a reminder, the transit technique is very, very simple. Physics is easy. Here is a star. Some of those stars will have planets and some of those planets will be orbiting at star such that the planet occasionally passes between us, the observers, and the background star. You look at the star long enough, you'll see it's mainly bright, bright, bright, bright, bright, as the planet moves in front of it, a dip. Transit technique. Very easy. That's the sort of data that you see. The radial velocity technique is somewhat different. Here you're looking at a star. You're looking at its spectral lines. You're looking at wavelength. You put the light through a spectrograph. You see absorption lines from the star. But an unseen planet will be maybe pulling that star backwards and forwards relative to you, and what you see is a Doppler shift. We're all happy about Doppler shift? Splendid. So what you see is this regular Doppler shift, you see over time the star being pulled away towards, away towards, away towards, away towards. We use neither of these techniques. That's easy physics. Let's get Einstein in the room. Come in, Mr. Einstein. Einstein's general relativity predicted that the gravitational field of a massive object would deflect light. Here's a cartoon showing exactly that. Here's us. This is the planet Earth on which we are all. Here is a background star. If we have a background star and a foreground object like another star, almost exactly the light from the background by the gravitational potential of this foreground object in a manner similar to an optical lens. Take a lens and pass it between your eye and a street lamp in a motion like this, and you'll see the light appear to get brighter than fainter. Pretty much the same deal here, except our lens is made of gravity. The very space, time fabric. This doesn't have to be a star. This is the cool thing. This thing here does not have to be a star. It could be, for instance, a black hole. Or we care about it as the gravity because we're not looking at the light coming from the foreground object. We're looking at the light coming from the background object. So this thing could be a star with gravity, a black hole with gravity, or even a planet by itself with no star. One of the primary recent discoveries made by the Michael Einstein community is that there is strong evidence for a large population of solivalent planets. Free-floating planets to everybody. I don't know why somebody came up with solivalent, but anyway. Planets without a star are floating through the galaxy. There's no reason, really, why we shouldn't think they don't exist. It's just because we can't see those sorts of planets with those other two techniques because we're relying on some kind of effect on its host star. The transit technique, the radial velocity technique needs a planet going around it to change the quality of the light coming from it. Microlensing, we don't care about that. But let's go back to the standard case where we think that the lens object is possibly containing one or more planets going around it. Those extra planets will change the characteristics of the lens. You take your nice, smooth, symmetric lens and then take a pickaxe and chip a bit out of it and you're going to get a glitch. So we're looking for these glitches due to the gravity of a planet. Another thing to distinguish the technique from the other techniques, the transit technique, direct imaging, radial velocity, is that those other techniques are looking at stars close to us, our local neighborhood. We, in contrast, this is the fried egg of our galaxy looking side on. This is us here about two thirds of the way out of the galactic disk, looking towards the very dense regions of the center of our galaxy at stars there, light being bent by a foreground object there, which is pretty much in the same direction towards the center of the galaxy. What we're getting at here is we're looking at different population of planets. We use this telescope of the 8-meter Moa Telescope from the start of New Zealand down here. This is beautiful Lake Tekapo. Tekapo townships are right about here. Mount Cook is somewhere up here, I'm not a geographer. And the telescope is right about there. Aerial view, this is where I spent a lot of my time staring at the sky in the freezing cold for one year. But then they built this nice new building here with much better facilities. This is that housing dome for the 1.8-meter telescope. This was built and designed by the Japanese and operates here. New Zealand company designed a particular part of the optics to install on there, and a New Zealand observer is there, full-time making observations. Here's what happens. The telescope with camera, comprised of 10 2K by 8K, sorry, 2K by 4K CCD chips. It's a large camera. It's observing about 2.2 square degrees on the sky. In context, that's large. We observe lots and lots and lots of stars. Each one of these stellar images is fed into a stack here, a cluster here down on the mountain. Curated by this man here, this is Professor Ian Bond. He works at NASA University up in Albany, and he makes pretty much all the computing work down on the mountain. The details of this machine here are, I haven't got that many details. There's about 10 CPUs running a particular type of image analysis. You're looking at effectively 50 million stars every night. 50 million spots of light, and you're looking for a change. So citizen science isn't going to help us in this case. However, Ian Bond developing a particular type of image analysis called difference imaging has made this routine. So, back to me. From the computer, we get all these data points. The two different colors there represent data from two different telescopes observing the same event. This is a micro-lensing event. I get these points. I put it through some code, which I'm using collaborators in Japan and also in the States, which is probably better commented than mine. I run this code in part on the HPC facilities here at Auckland. I get some results, and basically the long shot is I get a line short story rather, I get a line going through the data points. And that extra little bump here, which is contained in this set of data up here at the moment looks to be like a Saturn. A Saturn mass planet. That's what I do. Finding these sort of planets using these data. But you can see that there's a long path to go. Lots of people involved. Ashna may talk more about that in her talk. So, micro-lensing, this particular technique requires, like the other techniques, at least the transit technique, observing millions of stars. It requires delicacy in the image analysis and I again advertise the amazing efforts of Professor Ian Bond at Massey for doing this, making this routine. The whole project was set up. The model project was set up with the idea of finding dark matter. People know what dark matter is? Excellent. No, we don't know what that is. Would we like to know what it is? Yes, we would. Trick question. I can't claim it, but I saw it before. Somebody else is a brilliant one. Yes, we were looking for dark matter. There's a component of the universe of which we know nothing about. It was really a little embarrassing. So we call it dark matter. But the model project was set up amongst other projects around the world for evidence for dark matter because it's dark. We still think it has gravity. It's dark, but it has gravity. We can detect it through microlensing. That's the whole wheeze, right? We didn't need to get photons from it or any other kind of radiation from it to try and figure out if it was there and if so what it was. So microlensing. It's got gravity. Let's see if we can detect it. Did we find it? Well, we didn't find the dark matter so what do you do? Well, you repurpose your experiment. Let's find planets. So that's how we've evolved from a dark matter project. Well, I say we didn't find it. We set some pretty good upper limits on what it is. Great. Job done. No result. However, move on. Let's find planets. And these are our successes. What is interesting to me, however, and this is where I'm trying to haul you back to where we're supposed to be. The first scientific return from the MOA database, the MOA project rather, was not a confirmation of what is dark matter or not nor was it finding extrasolar planets. It was an emergent discovery. It was research into variable stars. You stare at millions of stars night after night. You can't help but noticing that some of those stars are variable. We know about these things but you have now got lots of them. Lots of data means you are able to make more fine questions about the topic of study. So while we're discovering extrasolar planets, what other discoveries await in the data? This is a question which I alluded to earlier on. So Alex will be talking about eclipsing binaries. Can we use this to find planets? It doesn't mean I don't know what he's talking about. I do because I taught him. What to research? But the idea here is he's classifying these data hopefully or lazily autonomously with the view of, okay, here's a piece of data. That's a variable star. Not too interested. It's one of those. Don't know what on earth these things are. This is where we write papers. This is where we get our research from. That's funny part of science. We know about these. I want a pile of weirdness to look at. It's remarkably easy. It is something of low hanging fruit with these databases. We set these databases up or rather we collect these into databases from experiments set up to look for one thing and we keep on plugging it trying to find that one thing because that's what we set up to do. We said we're going to find planets. Let's find some planets. But we're still shoveling data into our database and we haven't got the time nor resource nor et cetera, et cetera, et cetera slash funding to set people looking at those they are there for anyone to use. And this is a no question for you guys. You're the experts in this sort of stuff. And this year we have some friendly collaborators, the Ogil collaboration. They're doing pretty much the same science. They're looking for planets using microlensing but they are a Polish-US collaboration observing out of Chile. But they have put their data available via a website. Do we want to do the same thing? Is there a better way of doing it? I'm not sure. I'm not going to go into the details whether what they do is right or the best way possible. That's not the purpose of the talk. So we're going to do a little bit of a little bit of a review of these todays for making largest amounts of data available for people to use. So just to re-emphasize when you collect enough data you see weird stuff but everybody has different ideas on how to query the data. One of the things that I'm interested in doing is looking at the new technologies apart from the computer science side of things is a particular style of database which we should be looking at or using. And also the software here, first of all looking at open-source visualization of data it one of the tasks I did which was this low-hanging fruit what's in the database I took one of these databases and I asked a very simple question I came up with two statistics which were very quick to compute because my collaborator running their database wanted something very quick to compute because you had to run it on terabytes of data two simple statistics and out popped a result which was interesting very easy low-hanging fruit sort of stuff but I don't know whether it's necessarily the right way to go about it it just seems to be the right thing for me it was intuition or luck we don't know but that was my take on how to look at the data everybody else will have different ideas on how to query the data to make that easy for people to do so that kind of brings us in towards data visualization I like that to be an open-source project that had to be a piece of code which people if they wanted to could hack themselves rather than some proprietary piece of software which they have to obey the dictates of the programmers you don't understand how the code works at all that's what I'm interested in doing so there is some work here Ashraves and Paraview these are the ones to plug into the other for doing this sort of thing but it seems to be for quite a specific goal if you like I'm also interested in the interface between science and programming I read a talk by Gail Varoukwa and apologies for pronunciation but I was fascinated that the conclusion of that talk was that the science and the programming are interlaced at the data level so the data informs you of what science you can do and those two together inform you on how you program to get that science out so these things aren't disparate and I'm interested in looking at the machine learning paper from Psykit Python thing and the Python pipeline thing together to see whether we can utilize those things and what we want to do and just finally this is a plea or challenge or point I love Stellarium, it's a fantastic piece of open source software, love it poorly mentioned it in her talk what she's using what she's using Stellarium for we use it in our teaching laboratories as well to get students an idea of the nature of the night sky for those of us who live in cities and all we see is light pollution but I want to take it further and I thought when I saw this thing on the market I started drooling and it wasn't because I could play my favorite games because I'm not a gamer unfortunately but it was what can this thing do for data visualization I've been into data visualization you're in a room like this and you've got head tracking and by tracking what have you and you have 15 screens and you were immersed in the data and it's wonderful however I don't have that much money I don't have the necessary 100 kilo currency units to buy anything remotely like that but I might be able to convince the HDD to buy me one of these the idea being is I want to interrogate the data in a way which is intuitive and exciting and fun and stuff like that and what have you but that will be for me but I'm also quite happy interrogating data 2D on a computer screen I'm quite happy about deprojecting data you know 2D, 3D and back again and stuff like that I can see that but back to what I was talking before about computer science outreach if you like or utilizing all those fantastic you know graphics processing units between your ears all the school kids graphical processing units between their ears I want them looking at data in a way which will help me to produce science and I kind of think that this could be a way forward this could be something we could look at this is unfortunately owned by this so that makes me sad that's made me sad at all these are companies and this is how they work but I want to see what the open source community can do I want to see what the open source hardware community can do which will compete in a way which I don't have to pay quite as much as the full blown version challenge that's basic I believe research worldwide telescope because that has got Oculus Rift support splendid is the open source version of it I know that I started this by going hey can you not just like plug Stellarium into an Oculus Rift and you can fly through the space not that easy apparently I think it's doable there are some interesting things about human the human interface if you like on how you drive Stellarium through an Oculus Rift and that sort of thing because I looked up online and people have addressed this very thing and thought hey we can try and do this crash wall it's actually a little bit more difficult however I'm just here to say that there are applications for these sorts of technologies in the stuff I do in the classroom teaching students there's two aspects of this I grew up looking at Hubble Space Telescope images glorious images once they fixed it absolutely stunning astronomical images beautiful it's not enough for some kids these days some kids these days now what am I you take my point we are saturated now with sensational images and even the stuff which I certainly look at and go wow this really turned me on to astronomy even more than I already was it's fantastic some kids are still like a little bit blasé about it so we need to move in step with that use the latest technologies to basically impress them with the already impressive to me images that we already have for instance so this is going beyond just looking at data this is now engagement with students at a level with which they expect us to provide and also there are other teaching this goes to impact but also things like if you want to know what equatorial coordinates are for those of you who don't know this is how we map the night sky if you like we have a set of coordinates right ascension and declination they are essentially analogous to longitude and latitude here on earth but they are projected up into the sky so you can say right ascension 18 hours declaration minus 200 degrees you point to the telescope these are things you want to look at we try and teach this to students in our undergraduate lab for various reasons and this is what this means you say go read wikipedia wikipedia entry is correct but it's really hard to get a grip on it intuitively my view is you have something like that on your face you get it immediately so again this is trying to provide the best teaching environment that we can with the technology that we know we can do but I want to do this open source as best as possible that's a challenge of mine to the community so thank you very much for your time and happy to take questions thank you Nick that was wonderful if we have any questions we do have a few minutes before we have to head off for lunch who would like to start off there's a quick way wait sorry please wait silly what's the planet yeah what's that what silly what's it used for roaming planets oh Christ I don't use it because I prefer free floating but I'll try it I think any chance that those would add up to the missing don't matter no I'll try that hello sorry I said molecular hydrogen from behind me the plot that you showed of masses of planets that we've discovered they seem to cluster in groups is that because of the way we're looking for them correct that's right so the techniques of Doppler of a radial velocity technique of transits they started or large planets close into the host star that's basically the nature of the physics that's where they get the largest signal now you're getting better and better at finding planets further and further away because you're improving the data you've also got things like Kepler space telescope doing observing from space so that's starting to broaden out now microlensing weirdly is more sensitive to planets further away from its host star than the so-called Goldilocks zone which I really dislike the term but we understand it's that region where not too hot, not too cold microlensing finds planets in general further away out of that in itself is interesting because we still need to know how those planets form to get a good picture on how planetary systems in general form have you have you made use of the ipython notebook and for the data do you use that what is this thing that you speak of what is this ipython notebook it's for visualization of integrates like macplotlib and how they recommend checking it out if you're not aware of it this is why I came ipython notebook you discover a number of planets and different characteristics but you probably have an estimate of how easy it is to discover planets in different areas and you get an idea of how many planets there actually are because if it's 50 percent easier to get a planet from this area than there then you know the actual number of planets is going to be doubled over here are you asking when we're talking about regions of the galaxy for instance are you asking are we learning more about planets in different regions of the galaxy or is there any method like if you detect one percent of planets using this method yes you can run the numbers and get a population estimate yeah you can do that and this is where we sort of discover that actually planets aren't uncommon in fact they're the opposite they are littering the galaxy we assume that most stars actually have a planet of one flavor another again you can make also these predictions of these guys that we've discovered given the efficiencies of the telescope given the efficiencies of the observation program we can say well actually there's a big population of these as well so yes we can we run the numbers backwards given the efficiencies of how we work given the number of detections we work it backwards and we work out what the population should be if I wanted to get involved do you have an issue tracker for low hanging fruit how do I get in touch with you to help write some code tell you about technology the answer is no because we're not there yet again you would like to have the data ready for you to to look at presumably right just something that I can solve right no write that code that's the thing next time one of the attendees she's presently at Harvard I think there's a centre of rational physics but she is driving forward a project of let's have a competition who can write the best microlending code for instance right to model these things so that I suspect this is not necessary in her mind but I suspect we will be able to package and handle the community and say here's our webpage data you run it and tell us what you found that's what I would like to see so but we're all individuals I guess I'm using code from my colleagues have written and I'm adapting and using which is what happens you know when I left my company I had my code which I had written handed it on people took it and adapted it but we all end up running different sets of codes different coordinate systems and it's all very exciting when we try and analyse the same thing because we all got to agree on how we've set the problem up properly that has advantages and disadvantages the disadvantages you could probably see is that as you write your own code you understand it and unless you comment it very well and no one else understands it the advantage of having separate people running separate codes is that if you agree within error bars then you're more confident that you found the right thing if you're running one set of code which nobody understands or one person understands you're going to get the same answer usually put the same data in so there's a disadvantage there is this has been in the community for a long long time so there's tension there which we're trying to resolve and Jennifer Yee at Harvard is I think in some way trying to address that so we're not there yet but this is the sort of thing which we would like to be there yes I get the impression that a lot of the innovation and approach here is in the software and the methodology rather than the actual observation is that first of all that is a first statement and as such can you use your techniques on databases older databases and data that was collected before you started doing this work and make similar discoveries and mine it from there rather than making new observations as well I don't see why not and there is work in the community where people have gone oh here's a set of data that we have which is a contemporary database but let's go back and see what was found on those old photographic plays which have been painstakingly digitized and put online and stuff like that so people seem to go from here's something which I found an interesting contemporary data set but the information from previous data sets longer go say 10 years ago would be useful it seems to me and that's an impression of mine I don't necessarily say it's exactly always going to be the case but it seems that you discover something now with your contemporary data set and you look to see what else was known beforehand I don't know whether anybody has taken the data set from long ago and then gone forward or searched mainly I think because contemporary data sets are far richer than what we had in the past again personal impression rather than necessarily true okay thank you folks I know there's a few of you who still have questions and unfortunately we are heading into lunch and I think bellies are also rumbling thank you very much Nick that was a fantastic presentation very engaging I think you'll be around for a few minutes if anyone really has any burning desperate questions remember questions Nick will be around and we'll answer your questions now thank you