Subscribe
Share
Share
Embed
In this first episode of the R for the Rest of Us podcast, I speak with designers Cédric Scherer and Georgios Karamanis about how they learned data viz, why they use R for their work, and how they get inspired.
You may think of R as a tool for complex statistical analysis, but it's much more than that. From data visualization to efficient reporting, to improving your workflow, R can do it all. On this podcast, I talk with people about how they use R in unique and creative ways.
Hi. I'm David Keyes, and I run R for the rest of us. You may think of R as a tool for complex statistical analysis, but it's much more than that. From data visualization to efficient reporting to improving your workflow, R can do it all. On this podcast, I talk with people about how they use R in unique and creative ways.
David Keyes:Join me and learn how R can help you. In this episode, I talk with Cedric Scherer and Georgios Karamanis. They're independent designers who regularly produce novel and visually appealing data visualization. They came to prominence in the art community through their contributions to the tidy Tuesday social data project, and they've parlayed this into freelance work. I reach out to them after seeing a visually stunning date of his they made in 2021 for Scientific American on drought conditions in the United States.
David Keyes:The interview you're about to hear was done as part of my in progress book, R Without Statistics. We discuss how they got started with R and Dataviz, how ggplot and the tidyverse have helped them, how they find inspiration, and more. We don't cover code in this interview because discussing code doesn't really work well in audio. But the portion of my interview with Cedric and Georgios where we discuss their code is on YouTube, and I've added a link to that in the episode notes. Without further ado, I hope you enjoy my with Cedric Scherer and Georgios Karamanos.
David Keyes:Great. Well, thanks both, for joining. I appreciate you being here. I'd love to have you maybe first start out by just telling me who you are and a a bit about your background, how you got into R, and and what you use it for today. Cedric, why don't you go ahead and start?
Cedric Scherer:Yes. Thanks for the invitation. I'm Cedric Scherer. I'm based in Berlin, Germany. I'm originally ecologist or biologist becoming ecologist, becoming a computation ecologist.
Cedric Scherer:So I did my PhD on the focus of wildlife diseases and animal movement, and did a lot of our stuff. This is the usual tool you use as a scientist, and then got in touch with ggplot. And suddenly, my old passion of fonts and colors and design layouts and stuff matched with the programming, and that's how I ended up being mostly a freelance database designer doing most of the work still with r and ggplot.
David Keyes:Great. And Georgios, what about you?
Georgios Karamanis:Thanks for having us. So I'm a medical doctor, psychiatrist, and, I live in Sweden, in Uppsala, Sweden since 2016. I started the research about 3 years, so a little longer than that. 3 years ago, when my manager's supervisor recommended to download R as well so that I understand what our statistician was doing, and I got hooked up obviously immediately. And then I found out about the tidy Tuesday challenge and started doing database with vgplot.
David Keyes:It's interesting because both of you produce amazing data visualization, but, you know, it's not like you went to design school. Like, Cedric, you're an ecologist. Right? Do I get that? And, Georgios, you're a medical doctor.
David Keyes:You're a psychiatrist. Where does the DataViz stuff come from for each of you?
Cedric Scherer:I mean, data visualization is everywhere. Right? So for scientists, we have always 3, 5, 6 figures or more tables and stuff in our studies presentations. And I always had the passion and ideas of what looks good, what doesn't look good. I mean, it's a bit subjective, but there are obviously also some objective things people can agree on.
Cedric Scherer:And actually, I was thinking about studying either graphics design or biology and ended up using doing life sciences because you get paid better and you get better jobs. I'm not sure if it's really true, but that was my guess back then or what people told me. Data visualization is everywhere, and I'm pretty good in spotting problems. So visually, perceptually, I I would say. So also during my studies and then doing the PhD, I always when I saw presentation, I was like, yeah.
Cedric Scherer:Why don't they use more intuitive colors so better colors, colorblindness issues, and things like that, but also kind of like, yeah, why it doesn't need to be so complicated or just so black and white and boring? And, yeah, that's basically how I got into it. I thought like, okay, we could spark this up. I was also not thinking so much about data in ratio and all the stuff. I was just, when I started tidy Tuesday, it was just coming out.
Cedric Scherer:Like, everything you're not allowed to do in academia, I was just doing as I like, and that's actually also the fun part of one of the fun parts of 30 Tuesday, I'd say.
David Keyes:So for you, it sounds like doing the DataViz stuff was was a way to kind of scratch that creative itch that you weren't able to to get in some of the academic work. Georgios, were you using your were you using r during, like, your medical training? I know you said that it was a professor or someone who recommended it to you. I'm curious what the motivation for you to start using it for for the Data Viz, where that came from.
Georgios Karamanis:Frederick said that I was gonna say when I was looking at other people's charts or presentations. I had strong opinions about the designs part. But, I mean, I knew there was a thing, like, making charts, but not database. I didn't know about database until I started the coding in there. And then I think it was mainly through kind of Tuesday, I found out about database making, beautiful charts.
Georgios Karamanis:And I'm I'm using it for my research, of course, making Okay. I see. And,
David Keyes:Okay. So you use it for that as well, because,
Georgios Karamanis:I wasn't
David Keyes:sure. Okay. So it sounds like tidy Tuesday was a big pee I mean, that's how at least how I came to know the work that you do is it sounds like that was pretty instrumental for both of you. Is that is that accurate?
Cedric Scherer:Yes. So for me, I learned Digiplot before. So the the moment where I learned about was I always forget about 2015. I think it was 16 even. And I will we had to do a small multiples of very fitting to the topic today.
Cedric Scherer:And there were packages like letters, and I was doing it actually for my bachelor or so with bass r. So removing all the tick marks, removing all the stuff by hand, it was tedious, didn't look good. And I found about ggplot, and I was just pasting code from some tutorials like overflow. Nothing worked. The typical, like, aesthetics, I didn't know.
Cedric Scherer:Just get it to work somehow for the presentation. And afterwards, I was diving in. So I learned a lot. I spent tons of time, maybe too much on, yeah, really learning Digiplot and also, yeah, polishing my visualizations instead of writing my article. So that's basically how I found out that I want to do more of the graphics part and less of the, yeah, publishing, academia parts, basically.
Cedric Scherer:And then I waited, so I found out about Hey Tuesday. I was pretty inspired. I was also learning a lot already when not participating. So was founded in 2018 by Thomas, and I found out, yeah, exactly that time, I guess. And then one day after I handed in my thesis, I started kinda like working on TidyTuesday for several days in a row.
Cedric Scherer:So, basically, my vacation after my thesis was TidyTuesday. And I just as another anecdote, I just kinda remember exactly the moment when I saw Georgios first visualization. I was actually on a workshop, giving a workshop in somewhere in Germany. And I saw his I think it was the plastic one was your first one, right, Georgios? The plastics?
Georgios Karamanis:Maybe.
Cedric Scherer:Yeah. Yeah. So so he created this plastic spars out of bottles and stuff. So lots of icons and images. And I was like, who is he?
Cedric Scherer:He I haven't never even seen him before. Then I found out that you basically never used Digiplot before, basically, but you can tell that story. But, yeah, like, funny because I met so many people. Thanks. I mean, not just the followers, but we need people I chat with privately.
Cedric Scherer:And, yeah, it's a great initiative to learn, but also especially for the portfolio and the network part, I think it's just great.
David Keyes:What was the Georgios, do you wanna talk about the plastic visualization that Cedric was was mentioning?
Georgios Karamanis:Yeah. It it it was a data set about plastic pollution. And, I mean, it was, my first thoughts, and then the I mean, I struggled a lot in the beginning, of course. I was still learning gg12, and I had found, this package. I think it was the image.
Georgios Karamanis:So I took icons of, like, bottles, plastic bottles, PET bottles, and found out a way to rotate them randomly and make a a backlog. Yeah. And did
David Keyes:you do that? Was that in ggplot or what okay.
Georgios Karamanis:It was ggplot.
David Keyes:Because I know right. So because I know Cedric said, you know, he started when he when he started doing data Visitor, it was with with BaseR. Is that the same?
Cedric Scherer:There was no GGG back then. So when
David Keyes:when was when was that that then that you started, working with BaseR for making Data Viz?
Cedric Scherer:For me, the first time I mean, Data Viz, basically, I I learned how to use R for doing stats and LMEs and all the stuff, and then, yeah, then I was also doing plots, but this was the first contact I had with R was 2,008 that I really just followed of yeah. But was just really run and enter. I didn't understand anything of the code. I was not enough to get the math behind it. And then that I really was using it on my own, starting to use it was 2,012, I'd say.
Cedric Scherer:11, 12. But then yeah. Until tidyverse or DigiByte, it wasn't like, yeah, something you have to use because people use it. And I I actually preferred it over Excel, but was not really enjoying it in a way like I can spend the full day like I do it nowadays. Like, I can't stop.
Cedric Scherer:That's definitely not the situation before ggplot. That changed dramatically then.
David Keyes:Okay. Well, I wanna come back to that in a second, but I'm curious, Georgios, if you also did you did you use base r plotting, or did you were you always using ggplot?
Georgios Karamanis:I actually started because you need to use base on after handling the HD plotry, tidyverse, and everything because, I mean, everything was already there when I started coding in R. It was 3 years ago, so it felt much easier to do everything in gtplot from the beginning. So I don't know base are base are that good.
David Keyes:Yeah. Okay. Yeah. So, Cedric, you talked about how, now with ggplot, you're happy to work in it all day. Whereas with base r, you I mean, you made plots in it, but you didn't feel that.
David Keyes:Yeah. How how is it that a a package, an R package can have that much make that much of a difference in terms of your not only, like, what you can produce, but your enjoyment working with it.
Cedric Scherer:Yeah. I think there are 2 layers to that. So one thing is obviously the creative output. So it's not looking at significance levels and fittings and stuff, but I can really be creative about it. And maybe I'm not spending so much time on the code, but more on the right font and the colors, which is actually outside of R, but definitely tied to ggplot.
Cedric Scherer:So once you learn about ggplot, and I always try also to motivate that in my workshops. So even if you just grasp a few bits of it, if you get the general idea and clicks at some point, you're really free to create anything you like, basically. In a static way or some animated way, of course, but let's focus on this, like data visualization that gets into a report or into a manuscript, to a publication. You can basically create everything, and that's I know people I I once remember that one person was coming at me and saying like, yeah, you know, you're doing great stuff with ggplot, and I was create recreating all the ggplot stuff with face on. So it seems possible, but I don't, I hate it always OPA and PAR and all these settings, how to arrange things and super straight.
Cedric Scherer:So but I I I think it was also a bit hesitant to it to learn it because I'm not a programmer. I was, like, not really I was playing video games, but not really programming in my youth, and also didn't have a fancy computer at home, and I was studying biology and ecology, so I wasn't so much into coding, actually. And then, yeah, somehow, Jiggybot changed it, I think, from the design bit that I was kinda free and creative. And then came and really also enjoyed writing code because for me, it was way simpler to write. I know there's lots of discussion about it if it's better or not.
Cedric Scherer:I know both, but I actually never enjoyed using the apply functions, for example, and stuff like that. And if even if people now tell me, like, that's the same as you're doing with the per package might be, but it just feels more native to me and more intuitive to do it in the tidy context. Yep. Somehow then, it's part and also I had some some colleagues and friends who were basically showing it to me. So I also had kinda like this idea of getting as far as as they go while I wasn't really looking up to the base.
Cedric Scherer:People were like, oh, yeah, you do fancy things, but it felt felt like something new and hot.
Georgios Karamanis:Yeah. And
David Keyes:I'm curious. So Cedric, you talked about, you know, you've used Space Star, you talked about, it sounds like you did some Excel DataViz, but you didn't particularly enjoy that. I'm curious, Georgios, did you use or or have you used other tools to to do Data Viz?
Georgios Karamanis:I mean, no. The only thing I used, it was keynote to make some simple charts for my presentation, but not resolved.
David Keyes:That's interesting. So you went straight from not I mean, your your introduction to not just to coding, but to actually making data vis was r, it sounds like.
Georgios Karamanis:Yeah.
David Keyes:Because I I That's correct. Well, I mean, tell me if if you think this is right, but I feel like most people I meet have you know, like, my story, for example, I used Excel a lot and made, like, fine data viz in Excel. And I feel like moving to r has really helped me, especially ggplot. But but that's like but your story sounds unusual in that you you started doing DataViz in our did did was it something you just weren't interested in before or what? I'm curious.
David Keyes:Like, why all of a sudden 3 years ago were you like, oh my gosh. Like, I need to learn our do Data Viz. Where where did that come from?
Georgios Karamanis:I I I found out that Data Viz was a thing. I mean, I didn't know it before. So and then it's felt that everything came together very nicely because, I think, it's part part design, partly science. I mean, working with numbers. So I I haven't worked as a designer, but I I've always been interested in arts.
Georgios Karamanis:I've been a photographer before. I mean, I've read books about color and, composition, everything. So, that part was pretty easy for me. I mean, going into database. So it was Got it.
Georgios Karamanis:Only the coding part.
David Keyes:Oh, only the coding part, which for for many people is the the huge barrier.
Georgios Karamanis:So It wasn't easy, but
David Keyes:Yeah. Yeah. Well, I'm curious then. That sort of brings me to another question about both of you. You know, in addition, you clearly have strong technical skills now at this point that allow you to work with our you know, to make these high quality data visualizations.
David Keyes:But there's another side of it. Right? Like, you have to have a a a strong sense of aesthetics, understand kind of what, like, the design part of it is really important and understand what makes high quality data vis. So I'm curious. Georgios, you talked a bit about, you know, how always having well, you've been a photographer.
David Keyes:You've read books. Maybe, Cedric, I could throw it to you. Like, where did you develop those that other half of of the picture in terms of your design skills to bring those to be able to apply them to r and ggplot to make high quality data viz?
Cedric Scherer:Yeah. Interesting questions. I get this question from time to time. Also, kinda, like, can I give a brochure about it? And I always feel like, yeah.
Cedric Scherer:I'm not sure what to talk about. So what what I can definitely say that really this kinda like is not made up and I said, like, this kinda like is a child I had already this kinda like, you can call passion or mani maniac being, like, seeing wrong things. And I was also interested in fun, so I kinda like was a bit into graffiti and street art. Not really. I didn't enjoy it because I was still already a perfectionist, so I was mostly drawing on my desk, which I really enjoyed.
Cedric Scherer:I was using a lot of colors. I had a lot of markers with different colors. And I mean, this is fun already, right? Styling letters and getting proportions right and stuff like that. And also the layout and the composition in general, how much white space, how do you fill up things that look somehow empty.
Cedric Scherer:So I can't really pin it down to something. I just know that's kinda like something which I'm always into also with legal and stuff. I was always sorting them, and I think my legal is kind of or my buildings or whatever I was doing, creating, look kind of like yeah. Not too mixed, for example, so maybe a bit pedantic in my head. I wasn't I'm not really a reader that much, so I've definitely read about it, but more like whatever I need now.
Cedric Scherer:So if I kind of need to decide on colors, so I pick it up somewhere. Like, I learned from, I don't know, from Lisa, for example, about that. You should not use the stereotypic baby blue and, pink colors, or maybe you should, depends on what's your intention, but different topic. So I pick it up here and there, but it's not really like that I have the one book or the few things I just look up. And I think I'm kinda like learning a lot from just looking at other people's work.
Cedric Scherer:So that's maybe my kind of reading, looking at others' inspirational sources. And because I see the problems quite quickly or think I'm able to identify them, I can learn from these because I see like, oh, this doesn't work for me. Something is strange about it, so I'm not definitely not going to do that. So for example, if it's about alignments and stuff like that.
Georgios Karamanis:Mhmm.
David Keyes:What about you, Georgios? How how maybe if you wanna talk a bit more about how you developed your, you know, design skills and in building on that, if if people want to kind of, you know, develop a a similar or, you know, a strong sense of aesthetics as well. Do you have recommendations for people about how they might do that?
Georgios Karamanis:Yeah. I really I really don't know how I developed my design skills because, I mean, I'll I I've looked at design or art all my life, almost. So as Cedric said, I I kind of, it's quite easy for me to understand what's what works or not when I see something. And I've had no formal formal training, so I can't really translate it to sort of something more concrete. But I I I don't know.
Georgios Karamanis:I've been playing with design. I mean, drawing your logos, playing with colors and, I mean, drawing some stuff. So I've trained some stuff I've learned or seen, but other than that, it's it's hard to to pinpoint exactly.
David Keyes:Sure. I mean, what I've heard both of you say is that in a lot of ways, the best way that you have learned or that you've improved your design skills is is by watching and seeing what other people do. And I know that you have both, you know, been able to showcase what you do largely through Twitter. I mean, tell me if that's wrong, but but that's that's the sense that I have. Is has Twitter also been where is that also where you go to to see what other people are working on and get that inspiration, or are there other places that you go to?
Cedric Scherer:So, yeah, I definitely agree most of it showcasing on Twitter. I'm trying to also feed some other social media channels, pretty bad at it. So could be have an Instagram starting up, but works the best out of the other LinkedIn and stuff. They don't work well for me. So for me, it's Twitter because I mean, it's also different from the behavior.
Cedric Scherer:You can also post some work in progress while the others are more polished stuff. So that's maybe the interesting bit of Twitter to get inspiration. So I wouldn't say it's really about the design so much. So, of course, I take snapshots here and there are bookmarks when I see something, but I would say I see this more in newspapers or designers portfolio web pages or on Behance. Actually, I'm not so much on Behance anymore.
Cedric Scherer:I took quite a while from books. So I love really have books with. Yeah. It doesn't even need to be data visualization or the latest stuff or the old ones, or some others drawing. So could be also just nature walking around and seeing some, some combination of geometries or stuff.
Cedric Scherer:So really like the low level inspiration also. Yeah. I mean Especially for colors, of course, if you see some, I don't know, advertisement or some yeah, just some autumn colors or stuff like that. But, yeah, Twitter is is definitely very important. Not so much for inspiration, maybe on more for exchange also seeing, and maybe, yeah, the the work in progress stuff.
Cedric Scherer:So I think this is something which is for my feeling, at least, mostly shared on Twitter that people just, oh, I made this artful thing I didn't want to do. Right? This, how is it called accident by Arnon?
David Keyes:How is it called accident art?
Cedric Scherer:Yeah. Accidental art. Right. That's the one I searched for. And I'm posting this as well.
Cedric Scherer:So, I mean, I'm sometimes even polishing my accidents and then putting it out because I just like it how they kinda like the geometries and colors mixed up or stuff like that. So I think these are very inspirational things. So I mean, we also Garrison and I both did start, like, these having this history or the progression of our charts, the design. Uh-huh. And I really love this.
Cedric Scherer:So if people on or wherever kinda, like, have, for example, Dustin's, like, having a long blog post of their drawings and what were options they did and what did they not use. And often from these things they didn't use in the end, I see something like, okay, this could be something I want to try and recycle for something else because I like, how it looks like.
David Keyes:Yeah. Yeah. What about you, Georgios? Where do you go? Like, where are some specific places that you go to find inspiration?
Georgios Karamanis:Right. So mostly Twitter and, websites, and I have I follow some websites, RSS feeds.
David Keyes:Okay.
Georgios Karamanis:It's mostly yeah. Dataverse, of course, art artists of any kind like painters or comic artists or, yep. Typography. Yep. I Yeah.
Cedric Scherer:Yeah. Yeah. Yeah. Definitely.
David Keyes:When you get to a certain point, like, the the inspiration seems like it's no longer specific. Like, I just wanna see what data vis people are doing. Instead, you're looking much broader, and you're saying, I wanna see, you know, in nature, for example, or photographers, like, what they're doing and think about how what they're doing well can apply to my data viz work.
Cedric Scherer:Exactly. It's definitely such one. Yeah. Also on comics, I forgot about it. Comics, drawings.
Cedric Scherer:I mean, kids' books, for example, also super cool, actually. So, I mean, we both have have kids. So if I read a book and I mean, I just have a book laying here before it's super cute font. I love to use it. Some way I just need to find it, for example.
Cedric Scherer:Stuff like that. Yeah. Yeah. Or for me also, as mentioned, graffiti and street artists also. I mean, they are using extreme colors.
Cedric Scherer:And if you may notice with my work often well, at least in the beginning, I was using vibrant colors. And definitely often also using secondary and and more outlines, which is also coming more from the graffiti scene, I realized at some point. So there might be also some things I just were kinda taking over. But But at the same time, I'm still interested. I'm not doing it actively.
Cedric Scherer:I don't know the artist anymore, but if I go around in Berlin City, you definitely see lots of graffiti stuff, and that's definitely also some some sort of inspiration.
David Keyes:Interesting. Okay. But one last question. So, you know, there are any number of tools that people can use to make Data Viz. What do you think makes R particularly well suited, for doing high quality data visualization work?
David Keyes:But maybe Georgios, if you wanna start.
Georgios Karamanis:Yeah. For me, it's gzip log to and all the other packages, like extensions. It's very easy to do anything you want, and I think the grammar of graphics make sense. It's very logical and easy to work with. I would say it's also the
David Keyes:Well, tell me if this is a correct summation of what you're saying, but it sounds like what you're saying, it's the flexibility flexibility in terms of pack like, kind of extensions that people have made for ggplot. So you're like, oh, I wanna make a ridge plot. Someone's made a a package for that. But then also flexibility in terms of how you can output and share your work. Does that seem like a decent submission?
David Keyes:What about for you, Cedric? What what makes R particularly well suited for for Data Viz?
Cedric Scherer:Yeah. For me, there are there are really many, many different things. So, I mean, coming from academia and so on, the most important thing, at least in the beginning, was definitely the reproducibility and the kind of transparency and stuff, and also just being able to rerun the same thing again and not going to whatever tool you use. So I use PowerPoint and Paint in the beginning to change my labels when it wasn't there was no g g text, for example. Then there was the next request by your call for by your supervisor, by the reviewer, by the editor, and you do the same thing over and over and over again.
Cedric Scherer:So this is just annoying. So that's really what I like. And even though I have never written a full manuscript in R, like some people do it, I always kinda like like to have knitted reports of my analysis and the visuals. Then, I mean, for some context, for some of my clients, some of my work, you just need code based output. You can't do it manually because you have a pipeline.
Cedric Scherer:You have a workflow from data to the output, which is then online, a dashboard, or somewhere, the shiny web app or whatever. So there you need to be if you want to have something unusual, you need to be clever and also creative to get it work. And also then if you have changing data also to be kind of like clever enforcing any issues that could come up.
Georgios Karamanis:So So
Cedric Scherer:it's definitely flexibility and creativity. And then it's definitely also the community, I have to admit. I mean, tidy Tuesday 4 I mean, I was into DigiBot before, so I knew already that I might stick to DigiBot 2, but I know from many others that this is basically and not only a TidyTuesday community, but and this also changed a lot with Tidyverse, to be honest. So if you like or not like the Tidy idea or the Tidyverse idea, I think what really has changed and I really attribute a lot of it to the tidyverse people is the community and how helpful people are. So I'm always, if I have a problem with another program, it might be that I run into empty pages, old pages.
Cedric Scherer:I don't know whom to ask or the people want to charge me in in the our world. I mean, Google for everything, and you will find either a solution already. In the best case, there's a package or some code, or at least some person you might know or you might can contact and he's friendly or she, And that's really great.
David Keyes:Alright. Well, Cedric Georgios, thanks again for taking the time to chat and share the work that you've done here and all the database work that you do. I appreciate it.
Cedric Scherer:Yeah. Looking forward to everything you're doing with the others. Thanks for the invitation.
Georgios Karamanis:Yeah. Thanks for having us, and, it's been really cool.
David Keyes:Thanks again for listening. I hope you found this conversation interesting. If you have any feedback, I'd love to hear it. David@rfortherestofus.com. Thanks.