 Okay, so welcome again to our Meet the Experts program. My name is Tiffany Forman, and I'm a science educator at the National Center for Atmospheric Research in Boulder, Colorado. And I'm here today with my colleague, Nihon Cheroku, who is a project scientist with NCAR. And obviously we're coming to you from our own homes, but we're still doing our work and we wanna share it with you. So every other Thursday, we meet with somebody who works at NCAR and we learn about what they do in their jobs and we answer questions from those of you who are participating. One really cool part about working in a place like NCAR and why we love to do this program is because there are so many different types of jobs, such as being a scientist or an engineer or an electrician or a computer programmer or a safety expert or a machinist. We have people that do all of those jobs and more at NCAR and all those different jobs help support the scientific research that's happening here. So I'm gonna go ahead and turn it over to Nihon and he's gonna tell you a little bit more about what he does and then take your questions. Thank you, Tiffany. Let me share my screen. Okay, can you see the slide? Awesome. So yeah, welcome everybody and I'm glad you all could join us today. And I'll be talking about data visualizations and provide you with a bird's eye view of the kind of work that we do in the group at NCAR. So a little bit about myself, I'm Nihon and I'm a project scientist at NCAR. So I work in the visualization services and research group which sits under the computational and information systems lab at NCAR. And one of the primary things that I do at NCAR is to explore novel ways to visualize data. And these novel ways and applications that we develop are geared both towards scientific staff and also general audience to inform them about the science and the work that is done at NCAR. And before we begin a little bit of some background about myself. So one thing that you'll see while, when you hear from people, when you talk to them, like when you talk to other employees at NCAR, you'll see that most of the work that's done here is interdisciplinary in nature. And that's the case even with me. So in fact, I didn't start with data visualizations initially. So in my grad school, I studied LIDARs and environmental fluid mechanics. So a LIDAR is this instrument that you see over here. So it shoots laser pulses into the atmosphere. And then by looking at the light that is scattered back from dust particles in the atmosphere, a LIDAR can actually paint a picture of the wind field around the instrument. So in olden days, we could only take measurements that we call point measurements. We could take sparse measurements from different places. But with technological advancements like LIDAR, we now have an opportunity to get a much denser and richer view of the atmosphere around you. And this was the stuff that got me interested in data visualizations to begin with. So talk me about data visualization. What exactly is data visualization? So if you've seen, or if you've attended some of the other Meet the Expert talks, then you would have seen this picture on the left. It's about weather forecasting. So basically, if you are a scientist and if a scientist wants to simulate the weather around the globe, they would start by constructing a grid around the globe. And in each grid point, they would solve the mathematical equations. And the results of those equations are what will inform the weather forecasting. That's the basis of any weather forecast modeling. So once that is done, any output of the weather forecast model is essentially a bunch of numbers. So if you look at the image in the center, the output pretty much looks like this. So you would have a grid corresponding to the grid that is laid on the globe. And at each grid point, you would see some number corresponding to the quantity that's being measured. So in this example, it's surface temperature. And it's at this point that data visualization comes into play. So numbers by themselves are pretty hard to understand like what's happening. So the numbers by itself make it really hard for people to go through each cell and see what's happening. So data visualization allows you to see patterns in this data by visually representing these numbers. And the way we do that is we start with what we call as a color bar or think of it like a map legend. And you map all the values that occur in your simulation to this legend. And so every time you see a number, you color it with the color corresponding to that particular range. So in a way, if you look at the data visualization it's similar to coloring activity. It's like an adult's coloring book. But the only thing is that if you look at this visualization on the right, this is a practical visualization that we use. And it's created by Matt Raimi who also works advisor. You can see that instead of simple grids we have like millions of points. So obviously we can sit and color them manually and that's where we use computers. So we have computers and we write programs that sort of take this data and do that coloring activity for it. So if you zoom into one of this visualization you can see a grid pattern similar to the coloring book on the top. So that's in two dimensions. And when you go to three dimensions it's not very different. So instead of a single map or a single image you would stack these coloring sheets one on top of another. And instead of simple grids that you see over here you would have cubes. So a good analogy would be Minecraft. So if you've played Minecraft you construct things using cubes all around. So with data visualizations you could do the same thing except that instead of randomly placing these cubes you would have a computer program that takes in this data and colors that cube and places them accordingly. So one example of a 3D visualization as this visual of a hurricane. So moving on if you see that data visualization borrows techniques and tools from different disciplines and this is what I mean when I say data visualization as interdisciplinary. For instance we have science involved because you need to know what is being visualized and what is interesting for people to see. Secondly we need technology. So we deal with large amount of data so we can't do it manually and we write computer programs and we also create applications that help see these that help make the data visualization process easy and interesting. And secondly art also plays an important role because at the end of the day who wants to see a boring visualizations. So our does play a role in like making it reverting and keeping people engaged. So today we'll go through I'll touch upon and give you examples from each of these fields and it will be interactive. So as we move forward like you can enter some of the questions in the chat window below. So to start with art, one aspect in data visualization is choosing the right colors and this is the first activity. So and obviously you would have recognized the picture of Mona Lisa. So the question is what is strange in the picture on the left? So you can feel free to enter your answer in the chat. I'm seeing a couple of comments already. It's so red, lots of reds what people are commenting so far. That's a good observation. You see a certain color represented a lot more than others. Oh and the image is not as clear as somebody said and it's red where it would be lighter in the other one. Okay, that's a good observation. Just a few more seconds and we'll discuss. It's not as pleasant to look at. That is very important. Is the text play a big role in creating a good data visualization? So one thing that is different between these two images is that so these pictures are created by recoloring the Mona Lisa painting and you recolor using the color bar at the bottom. Remember the legend that I mentioned a couple of slides ago and if you look at the colors for the picture on the left, you can see that the colors sort of fade and blend into one another whereas the color bar on the left, these colors, they don't necessarily blend in. So they jump across all, they jump all over the place. And in fact, this is the notorious rainbow color map which you find very common and I've used it in the past too. So the issue with rainbow color maps is that like when you have colors that jump from one color to another, you can start seeing patterns that don't exist. So in this case, you've already seen the Mona Lisa painting. So you know what the ground truth looks like. So you can tell that by using a rainbow color map, you get a picture that is not quite right. However, when you're working with data, you don't know beforehand how a data looks like. So that's why it becomes important to choose the right color map because if you don't do that, you can end up seeing patterns that don't exist. For instance, this is one example that's often cited in papers. So if you look at this image, the first thing that jumps off at least for me is that demarcation that goes right through the center of United States from Texas all the way to the north. By looking at this picture, it feels as if there's a clear line that separates the Eastern United States and Western United States. However, if you look at the numbers, the numbers flow uniformly. The demarcation that you see is an artifact that comes from the colors that we choose. So this is the importance of choosing. Like you have the artistic liberty to choose any color that you want. However, you have to make sure that the colors are consistent and they flow from one thing to another. And another important consideration is that for people who are colorblind, if you use some combinations of reds and greens, then the visualization pretty much becomes useless for them because reds and greens would appear the same way. So, and that's the importance of choosing the right colors. Now, moving on to science. So one thing that is important is something called map projections. So map projection is a way in which you take a sphere. So earth is a sphere. And you take anything that's on the surface of a sphere and you laid flat on a flat surface. And it turns out there's more than one way to represent a globe on a flat sheet of paper. And these are some of the visualizations that were created by Matt and Tim from, again, Weiser. And it turns out there are a lot of these map projections and this is another design choice that you have when you're showcasing data that spans a globe. And the reason for having so many of them is that there is no such thing as a perfect map. So when you take a sphere and flatten it, you invariably stretch out certain areas. And the stretching process distorts certain regions on a map. So based on what you're trying to visualize through your data visualization, certain map projections work better than the other. In fact, who knows, like maybe in the future you might have to create your own map projection that represents a certain quantity that you are interested in seeing. So it's an open problem and also something interesting. And when I speak about distortions, here's a quick activity for you guys to try out. So I'm from India and let us say I want to move from Boulder, Colorado. I want to travel and visit my parents who live in India. And I obviously want to take the shortest path from Boulder to India. So assuming I want to take the shortest path, which one would I choose? Would I take path A, path B, or would I take path C? So you can enter your answer in the chat below. I will reset the timer. And if you need a hint, the animation in the lower right corner shows you how this map was obtained from a spherical earth. We have some guesses for C so far. Great. Anybody else want to take the guess? There's another vote for C. Come on, does anybody want to guess A? Come on. A is feeling lonely. Someone said looks like A from this view. That looks like the simple answer, right? Great, thank you. What it turns out, C is the right answer. And you can see that through this application. So this is another application that we developed that helps people try out different map projections. And the reason why C is the shortest path is that this map, you need to put it back on a sphere. And when you put it back on a sphere, C turns out to be the shortest path on that curved surface. And the projection that we've used, it's called the equirectangular projection, or cylindrical equidistant projection. And this type of projection is good for visualizing data near the equator. However, if you, let us say you are a scientist who was studying CIs, and you want to see CIs that mostly exists at the pole, then this visualization is probably not an ideal thing to see because it distorts the land near the poles. And if you want to visualize polar data, then you might probably choose one of these projections, which basically looks at the earth from the north and south pole. So with this projection, there's a lot of distortion near the equator and none at the poles. So all this to say, you need to choose the right visualization that goes with the science that you are studying. And this also transitions us into the next aspect of visualizations, that is technology. For instance, this application that I've shown uses something called augmented reality. So augmented reality is a technology that allows you to see digital objects and place them in real world. Now, before we go dive deeper into it, here's another activity. So this is an example visualization that I've created during my grad school. In the center is Pokemon Go, and that's an example of an augmented reality app which is very popular. And I think you must have seen it one way or the other. And the video on the right shows one application that helps visualize some of the data. So the other question is, what are the similarities that you see between the Pokemon Go screenshot and the data visualization application on the right? And bonus points, if you can find the Easter egg, you can enter your answer in the chat below. One suggestion of 3D. Yes, data visualization is getting, is increasingly three-dimensional in nature. Here's, I guess, is the Easter egg, a Pokeball in the one on the right? I did not observe it. So maybe there are two Easter eggs. And someone else says they are both on a real background. That is an excellent observation. A few more seconds and we'll move on. So as someone has pointed out, it is true. So both of them use something called augmented reality. And if you look at the Pokemon Go, you have this animated character that is placed in the real world. And this is primarily used for entertainment. But we could use the same technology that is used to create games like Pokemon Go to visualize data. In fact, the visualization that you see over here with the wind field flowing around the crater is not very different than the animated character that you see in the center. And like using techniques like augmented reality, you can bring the data that was initially collected from the real world and you can visualize it in the real world. So traditional visualization revolved us having to visualize this stuff on a two-dimensional computer screen. However, using techniques like AR, we can then bring this data and overlay it on at the location where the data is collected. So it makes the interpretation process more intuitive and engaging. And now for the Easter egg. If you see after giving all the spiel about using the right color maps and not having to use rainbow maps, I've used rainbow color maps too in the past. And the reason for that is at that time I didn't know and you learn over time. So yeah, always one thing that really helped me is that there's always more to learn and you learn and grow as you move forward. So yeah, don't use rainbow color maps. So that's a bad example. So moving on, so we have created some of the, like some other augmented reality applications too. So one advantage of augmented reality is that, as I said, you can see digital objects in your real world. And you can also scale those digital objects to their real size. For instance, if you look at the pictures with the pets on the left, the white thing that you see is a representation of a hailstone from that fell in Vivian, South Dakota a couple of years ago. And that's the largest hailstone that was ever recorded in the United States. So you can now, you have an opportunity to place this hailstone in your surroundings and see and compare it with some of the commonly found objects around you. So to, and this is a web-based application. So the QR codes take you to that respective application. So you just need a smartphone. It works on both Android and iOS. And we have some of the other models too. You can place a supercomputer in your house to see like how big that is or place a hurricane in it. And the application to the right is another example of where we use augmented reality to engage general audience and inform them about the visualizations that we produced in our group. So Meteo AR is, I think the simplest way to describe us, it is an interactive science sheets. So the science sheets have information about different data sets that we've worked on in the past. And the QR code, which is over here slightly hidden by the virtual object, it informs the application to overlay a certain data set corresponding to the page over there. And by turning the page, you can change the data set and it gives you an interactive and fun way to explore different data sets. So going into the future, so what's next? Yeah, we do continue, we continue this work even now, like we are on the lookout for novel applications and how we can make visualizations better. And one of the underexplored area right now that we are interested in getting into is accessibility. So one theme that is common in most of the stuff that I've described earlier is that it's all visual in nature. And that poses a challenge for people who are blind or vision impaired. So by bringing in the technology and techniques that have been used to make certain things accessible to people who are blind and vision impaired, we'll be able to make visualizations accessible to people who, he couldn't access these visualizations in the past. And we are at the right intersection to do something like this because technologies like augmented reality, as I've said, they help you bring the digital objects into your real world. So what this means is that in the past, when you had to work through a two-dimensional screen to visualize data, you can now visualize it in your 3D space. And that makes it intuitive for people who are sighted and also it gives people who are blind an opportunity to find novel ways to walk through and explore data sets. So the image on the top shows like how your phone running an augmented reality application makes sense of the world around you. And this is my apartment. And the yellow dots are basically what my phone thought were interesting features in the world. So it could be a texture or it could be something on the carpet. Or like over here, you can see it's the couch. So it's taken patterns from the couch. So by using this information, a phone is now aware of like where it is in the surroundings. And then you can start placing relevant information and you can start placing visuals in that three-dimensional space that could help inform people who are blind and vision impaired on how to explore that space. So moving, so going into the future, like one thing that I would summarize is that like most of the work that I do is interdisciplinary nature. So if you are interested in a certain field, be thinking about how that can apply to some other problems. So there are many problems, but you'll be surprised at how you come across solutions. Like maybe the solution to your problem is not in your field, but you might get it from some other field. So yeah, with that, I will take further questions. And I thank everyone for being patient and attending the stock. Great. Thank you, Nihant. It's pretty awesome, the work that you do. And I'd love to take any questions to share with Nihant. If anybody has any questions about what you've heard or you wonder something else about the work he does, I have a question, Nihant, while other people are entering theirs. With the choosing your colors, are there kind of known color palettes that work better for certain applications? Or is it kind of just using your eye and figuring out how to best replicate what you're trying to do? So now you do have a lot of, as I've said, like artistic liberty to use the kind of colors that you want to use. However, people often tend to use the colors that closely represent the quantity that's being visualized. For instance, if it's a visualization about sea eyes, maybe a palette involving whites and blues would look more appealing and it's more convenient. And at the same time, if you're visualizing temperature, it's common to use any color palette that involves reds and blues, preferably blues for colder temperatures and red for warmer temperatures, because that's what we are used to expecting. Like, yeah, red things are hot and cool things are blue. So you can base something off, like some of the previous understanding that we have about the phenomenon that's being visualized. But again, having said that, you can use some other data sets, some other colors too, just making sure that they flow smoothly one into the other. Interesting. Got a couple of questions coming in and also just an appreciation for how cool it is that you're working to make all of this accessible, to visually impaired and blind communities. So that's a great application for this. Yeah, that could be done in that field and honestly, just now getting into it. So we're still at the beginning stages, I would say. That's great. Somebody else wonders, how did you know you wanted to go into this field of work? Well, the answer is I didn't know from the beginning. It was like, my work informed me like where I could go next. So like one thing that I see is that I was in general interested in learning about new things. So as in when I found different interesting things from some of the other fields, I tried to apply them to my field and I let my work sort of dictate where it takes me next. So like I wouldn't, it wouldn't be accurate to say that I knew right from the beginning what I wanted to do. Great. Here's another good question. It seems like a lot of scientific visualizations require complex coding. Are there any programs or tips that you'd suggest for newbies? So the good news, so two things. Yes, it is true that data visualizations can be complex, but at the same time, the good news is there are many tools that are available online. So depending on how much experience you have with computers and stuff, like this is almost everyone can use computers. So you might be able to do data visualizations without knowing to do any coding because there are applications that, in fact, like there are some applications that we've developed at NCAR, some of the other groups too. So we have something called NCAR command line, like language to NCL and vapor. So these two are softwares that we've developed and if you are using these softwares to visualize data, you would have a graphical user interface that you can use to load the data and change the colors and play with it and do the visualization without getting into the coding aspect of things. So yeah, look for data visualization tools and applications that are already out there. Great, thank you. Sure. And somebody else asked if you can say a little bit more about the accessibility piece of this, is there anything in use now around that? Yes, so basically, excuse me. So we are using the same technology that allows you to place these digital objects in like 2Ds, like in your 3D world, just like a visual hail model. We are exploring, augmenting the real world with audio cues. So in fact, we are working with a major museum in trying to make the indoor spaces accessible and to help people who are blind and vision impaired navigate those faces. For instance, if you look at the image in the lower right corner, so you have a bunch of posters on the wall and by using an application, so the user over here, she's holding a phone that's running an augmented reality application. And when the phone intersects a certain area or when it comes in close proximity to a certain map, or sorry, a certain poster, it can read out what's there on the poster. So in a way, the idea is to give people who are blind an opportunity to explore the surroundings like how people with vision do, in the sense that you can see what's around you and you can decide whether you want to go deeper into that particular topic or move on to another topic. So this is our work in progress. So we have one prototype and we are working with another museum to scale it up and to create an application that can be widely used by other organizations too. That is just fascinating. It seems like there's so much that can be done with that and will be as time goes on. I don't see any other questions. We'll make a last call for questions and there's a lot of appreciation in the chat. A lot of awesome presentation comments. All right, Ben. So I would just like to say thank you, Mihan, for joining us and for telling us about your work a little bit more. There is one more question. If we're a little bit over time but if you've got enough time. Absolutely. This is a really good question. What classes did you enjoy as a kid? I actually enjoyed my art classes. I mean, there's that piece. Yeah, and I know there's this tool online called Google Sketchup and it allows you to construct these virtual worlds online. It's a, I think it's a web-based tool. At least it was a web-based tool like many years ago. And things like modeling, like 3D modeling in which like you're given a bunch of options and you can construct things together, those activities. And my favorite class was like obviously science and that's what led me to choose a science-based major going into the future. But it was mostly science and art and I had to connect the dots like after I moved forward. There are a lot of connections though. That's what's so fascinating. Science and art really are connected in so many ways that we don't often think about really. Yes. That was a great question. So again, I would just like to say thanks for joining us. It's been really fun to hear about what you do and thanks for sharing that. And thanks to everybody who has participated and joined us today. For future reference, we do meet the experts sessions every other Thursday. So hopefully some of you will join us again in the future. The next session is on November 12th and that's going to be talking to an aircraft mechanic who works for NCAR's research aviation facility. And he's gonna tell us about how the airplanes are used and how important his work is in keeping them ready to fly and ready to go on research missions. So we'll post the link in the chat to our, I'll do that right now to our meet the experts page where you can follow our schedule and you can find recordings of all of our past meet the experts sessions. That's the link to that. And then I'm also gonna share the link Nihant gave us a link to their visualization gallery of Nihant's work group. I'm gonna put that there. If you'd like to check out more with the augmented reality and the things that they're doing. And with that, I think we'll close for the day. If there is anybody still here who's in grades five through 12, if you wouldn't mind filling out that quick survey, we would really appreciate it. And I'm going to, it looks like Tim's gonna paste the link to that survey. It is outside of this presentation. If you don't mind going to that page for the survey we would really appreciate it. And thanks everybody for joining us. Hopefully we'll see you again soon. Thanks everyone. Bye. And then take care. Bye.