Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Episode #127: OpenVisConf Recap
164 Plays7 years ago
Welcome back to the podcast! On this week’s episode, I sit down with three amazing people from different parts of the data visualization world to recap the 2018 OpenVisConf conference held in Paris, France earlier this month. Who do I...
The post Episode #127: OpenVisConf Recap appeared first on PolicyViz.
Recommended
Transcript
Introduction and Conference Overview
00:00:11
Speaker
Welcome back to the PolicyViz podcast. I'm your host, John Schwabisch. On this week's episode, I am sitting in Paris, doing a recap of the 2018 OpenViz conference. We're in Paris, which was amazing. You're hearing a few people in the background still chatting. And I'm fortunate enough to be surrounded by three amazing artists.
00:00:31
Speaker
data visualization designers who wanted to sit down and chat about their favorite parts of the conference. So what we're going to do is have each of them introduce themselves for you folks who are listening at home. And then we're going to talk about our favorite talks and things that we took away from. So Mike, I'm going to start with you. All right, excited to be here. My name is Mike Freeman. You can find me on Twitter as at mf underscore viz. And I'm a faculty member at the University of Washington Information School.
00:01:00
Speaker
Hi, I'm Zan Armstrong. I'm Zan Strong on Twitter. I'm a freelance data visualization designer and engineer. I'm currently working on a long-term project with a team of scientists at Google. Hi, I'm Nadie Braemer. I'm a data visualization freelancer. I'm in the name of Visual Cinnamon. And I work at home with my cat in Ellens. Terrific. Great. Well, welcome all, Zan and Nadia, back to the show. Mike, this will be your first time, but we'll get your back soon. OK.
Impressions of OpenViz 2018
00:01:30
Speaker
So, two days, first things first, overall impressions of the conference. Amazing. Yeah.
00:01:39
Speaker
It's really amazing, really special to be here in Paris for the first time in a conference. A lot of Americans still came over because they loved the comics in the past and got a ton of Europeans and other people around the world, so a really special community. Yeah, exactly. I thought that was a really good mix of the multi-missionality this time, especially. And also the talks, I really love just the general balance of more design-focused, more math-focused, more team-focused, really, really good comics. Yeah.
00:02:08
Speaker
Great, I also want to talk about a little bit we should talk about like the structure of the conference because I think there's there's like a lot of time for Talking and eating lunch and like networking that a lot of conferences don't have and I feel like that's one of the things that makes this conference special
00:02:22
Speaker
Okay, so we're gonna start with you, Mike. Your favorite presentation in the last two days, I'll of course link to the open this site on the show notes.
Highlighting Key Presentations
00:02:30
Speaker
And soon, they will also post videos. But I think we're gonna have at least three or four of our favorite talks from last two days. So, Mike, let me start with you. What was your favorite? What did you like about it? What did you learn? So, one of my favorite parts about
00:02:46
Speaker
been not only a focus on developing our own methods about design or programming techniques, but there's been a lot of
00:02:56
Speaker
making really careful design decisions because of the impact they'll have on other people, and also taking these skills and bringing them to bear on important political topics. So, not surprisingly, my favorite talk came from Aaron Williams from the Washington Post, whose talk largely centered around an article that they have released called, America is more diverse than ever, but still segregated.
00:03:17
Speaker
This is a really thoughtful use of various metrics of segregation to try and understand the racial composition of a variety of cities in the United States. It was a really beautifully done piece and incredibly delivered talk and was a pertinent example of how we can take data visualization skills and help shed light on nuanced and emerging topics in the U.S. and around the world. That's great. Okay, so, Zan. Zan, favorite talk?
00:03:47
Speaker
I'm going to cheat a little bit and say one thing that I really liked about two talks was I really enjoyed
00:03:54
Speaker
both break the rules for a purpose and so Matthew Kay talked about Bayesian analysis and how it's really important to make it designed to make it hard to ignore uncertainty so you can't just like not look at it and that really influences design and making he pointed out the New York Times anxious thing that we all know about the anxiety is actually like that's here some of the uncertainty that can be a positive and I also really enjoyed
00:04:29
Speaker
in a slowness of interpretation and engaging was an intentional part of that presentation that I thought was really cool.
00:04:36
Speaker
Yeah, my friend, sorry, that was a tough choice, but eventually I think I really, really liked Steve's brand, Clinarium's brand, from Northwestern University, who was talking about thinking with data visualizations fast and slow. And this was really about the research that he's been doing about how people interpret visualizations and how some things are, we can sort of see at a glance, and some things take more time to process and hit some. I really liked that talk, especially, I've been into the kind of
00:05:05
Speaker
we've seen, how we've not really seen reality for ever since I discovered data visualization. And he really brought it home towards data visualization. So he had amazing examples. Something that we see like an outlier immediately, but make a slight change of it and suddenly we have to do a lot of sort of slow processing to do that. And I really like how we showed his research in very approachable ways and the things that he found and things that he wanted to do in the future. So yeah, that was really, really good to talk about.
00:05:35
Speaker
Those were all great.
Visualizing Uncertainty
00:05:38
Speaker
One theme, I think, that at least I saw that sort of flowed through a lot of the talks was on uncertainty, which as Andy mentioned. Also, of course, with any academic research, we've got a lot of uncertainty going on. Before we talk more about the conference, which I want to do, I want to get your thoughts on visualizing uncertainty. I think it's like one of the biggest challenges. And some of the talks were about
00:05:59
Speaker
you know, Bayesian inference and showing, I would say, like uncertainty in sort of complicated looking ways. But how do you get to present data with uncertainty to...
00:06:11
Speaker
you know, everyday people who are looking at a bar chart or looking at a line chart. So anyone have thoughts on like how to visualize uncertainty, how to convey uncertainty, statistics, probability to our audiences? I don't have any of this. I want to point out that something that is particularly complicated about it isn't just about getting three times as many numbers. It isn't just about showing now in upper and lower bounds on top of the average that you've calculated. But it's also balancing
00:06:42
Speaker
generating a belief in your numbers, right? These are the right numbers that this is clear in a sort of authoritative way and at the same time communicating they might be wrong. I think a lot of people want to both express their confidence in their methods or in their approach and at the same time express that those approaches inherently have some sort of limitations.
00:07:04
Speaker
and striking that balance makes it additionally hard. So I'm sure one of these brilliant people here has a great way to do that. So yeah, I think uncertainty is a really big deal. I think especially around the question of where the data comes from in the first place and what data isn't represented. Is your data actually representing what you think it is? I think so many times there is a talk
00:07:29
Speaker
who was, I think Mariah was talking about a lot of proxies and that we often use proxies and those can be valuable but we should be really intentional about noticing the proxies that we're using and recognizing the potential challenges there. And I think also there was a number of talks about machine learning here which I think is also an interesting place for uncertainty that there's kind of inherently a lot of uncertainty in how the decisions are being made.
00:07:58
Speaker
people explore kind of exposing, exposing some of that and also like using that uncertainty or being, accepting the uncertainty and still using it as a tool and how do we think about that? Right. So do you think about communicating the uncertainty to people in specific ways or is it more for use and I guess more of I'm working with these data. I want to make sure I have the data right and I'm going to present it, but maybe not actually, you know, showing the uncertainty or showing the, you know, confidence mountains or something like that.
00:08:27
Speaker
I know, it's a hard question, right? Yeah, this is a really hard question. I mean, I want to say, like, yeah, of course I'm going to show the uncertainty, but I actually... I mean, I think it's really important to push in that direction of showing the uncertainty.
00:08:41
Speaker
It's really complex in terms of should that, are we showing that in the photograph? If we are, how do we show uncertainty about uncertainty or what it means? There was an academic paper a couple of years ago about all the different definitions of uncertainty and using this very similar glyphs for very different definitions of uncertainty. So I think it's, if you see it more, just a constant of,
00:09:21
Speaker
Yeah, I have to admit, I don't have a real good answer to that either.
00:09:25
Speaker
I haven't had to, or maybe have ignored in certain cases for a long time. I do have something that is kind of related to that, and it has to do with aggravation. So I do think a lot about the mean or the moral of the 50% point. And that it is not necessarily usually not a good sort of way to show the underlying.
00:09:47
Speaker
distribution so I'm always things that I like to do in my visualizations if possible is show the context so show a lower level of detail without having to accurate so that I can tell people more I can maybe highlight the median but then I still want to show the other things that are surrounding it where it comes where that median comes from
00:10:06
Speaker
which always in my mind these uncertainty and kind of these aggregation things are kind of connected in there, both difficult to do but the one thing I think about a lot, whereas the other thing I have to... That's okay, that's okay. It's an interesting, it's an interesting topic.
Subjectivity in Data Visualization
00:10:26
Speaker
So the one talk that I think was sort of uniformly beloved last two days here was the talk by Amanda Cox and Kevin Cooley from The Times.
00:10:36
Speaker
Who wants to summarize the talk? So their talk was called Disagreements and it was a wonderful back and forth between the two of them on things that they just generally don't agree on, which I thought was really great. It's not that everything is always
00:10:54
Speaker
good or within officialization there's always a sort of a subjective factor to it and that's what they really showed in this talk and they had their own reasons for why maybe Amanda is okay with charts not being simple whereas Kevin is like both charts
00:11:11
Speaker
to be simple. That was just one of them, and then they showed some wonderful examples, and they showed some wonderful personal Slack messages. And it was a really good set of these examples. I mean, you think about just that, it's a gray area, and there are reasons for using one or the other in different circumstances.
00:11:37
Speaker
Right. Okay, so do you find yourself on Team Amanda or Team Kevin? I can be very clear that I'm on Team Amanda. Okay, okay. Okay, Team Amanda. So, Zan, Team Kevin or Team Amanda? Team Amanda. Okay, so why Team Amanda?
00:11:56
Speaker
I really enjoy how she thinks about things. I mean, I should say, they're both amazing, and I super love, so I just love good perspectives. Yeah, I enjoy how she frames things. I think I really enjoy, she sees complexity even behind the simple, and I think even simple things are complex. And I enjoyed how one of the questions was with that, the famous odometer of the election, is that a simple chart or is that a complex chart?
00:12:26
Speaker
And Kevin said it was simple, and Amanda said it's complex. And I think there's a lot of complexity behind simple things that end up kind of simple in the end. And I think that's why I'm a fan. Yeah, I think that was one of the great examples of like, is that a complex chart or a simple chart? I think it's a super complex chart. So I'm definitely up for that one. I'm on Team Amanda. OK, Mike, Team Amanda, or Team Kevin?
00:12:50
Speaker
follow suit. Not to say team Amanda, but with my former question, which is to avoid answering this question directly and say, what I thought was so amazing was not one side versus the other goal, was the generating source of these different things they were in intention about, was because they wanted to figure out
00:13:12
Speaker
the best way to do something for their audience. It wasn't just about a whimsical opinion about X or Y or Z, or that there was a preference in color, but it was such attention to detail and such care.
00:13:23
Speaker
or the consequences of their actions that there was beautifully and hilariously debated. But the reason that there was such fervent debate is because there was such concern about what sort of interpretations or misinterpretations there would be. So I think the thing that I decided with was not one person or the other, but the tension between them. But I thought I was pushing their workflow. Yeah. Yeah. OK. All right. So let's talk about the structure of the combat conference. So we've all met at multiple conferences, many conferences.
00:13:53
Speaker
What, in your mind, makes this conference sort of special in terms of the structure? And then I want to talk about sort of take-home lessons from the two days, but just in terms of how the talks are put together, the breaks are put together, what does this conference stand out in a way that's special relative to other conferences? And if you think it's a great way to structure it, should other conferences try to mimic this approach?
Conference Format and Networking
00:14:18
Speaker
I mean one simple thing that I really love is that it's a one-track conference and everything is videotaped and put online later and I think that's both of those really and there's a lot of care taken and really thinking about how to curate that because everybody's going to be sitting and watching every person and that's going to be a long-term experience that hundreds of thousands of people will watch later. I think that really people are really proud of what they present. They put a lot of work into what they're sharing and how they're sharing it and
00:14:47
Speaker
really views our audience's time, both here and everywhere. And I think it's just a really thoughtful format that just creates really fantastic talks.
00:15:00
Speaker
As you mentioned they put a lot of emphasis on having sufficient times for breaks and for lunch and they put the structural pieces in place to make sure that people had time to share ideas with one another that it wasn't just a passive receiving of the information through the talks but also you had time to on a two-hour boat ride around Paris spend time with
00:15:22
Speaker
the other people in your field or in other fields and discuss these ideas and similarly do that through your coffee breaks and lunch and providing that amount of time and those fun environments in which to do that makes it as true as to say all of us were able to do one another.
00:15:37
Speaker
Anything else to add? No, they said it all. All right, cool. So what are there specific things or maybe just general themes that you are going to be able to take home with you and use in your work? Either specific techniques or coding snippets that you saw or sort of more general themes that you saw that you learned that you think you can use when you go home?
Incorporating New Techniques
00:16:04
Speaker
One very practical thing for me was the t-snee clusterings. That was in the first three talks, so I won't show the original section that involved this kind of clustering technique. And more of its cells have showed how you can then go through the grid-based thing. And Ian was talking about the quickdraw.
00:16:21
Speaker
different kinds of people drawing different kinds of cats and how you can visualize that and I really want to use it. I've used it before but now I have some other new ideas and I've already talked with Ian about some of the things that you can help me with. Dragons and roommates.
00:16:38
Speaker
Wait, now I can't ask you about dragging a mermaid project? Okay, we'll hold on to that one, okay. But that was a very practical one. And the other one is a cat. That you can do more than you think you can. And that these people that are showing these techniques and that the episode is truly amazing and then they have like a tidbit of information. I was using this or I used this hat. This is a general idea of
00:17:04
Speaker
If we're all just people doing our thing, and it all starts from a really simple starting point, and then having that spark of an idea that's really good, I guess that's sort of my general vibe that I got here. So you feel kind of inspired to go back? Yes. Okay. I mean, that's a huge part of it, right? Okay, great. Yeah, yeah, yeah.
00:17:24
Speaker
So I feel a little repetitive. Something that's been on my mind is how do you, I think we often think of your statistics and there's this and like the complementary, but like they're kind of different things. And I think that there's a lot of interesting things going on, whether it's ML or sampling or other other places. There were some nice examples, Ian Johnson, Moritz, Sean Carter.
00:17:47
Speaker
and probably a few others that I'm missing about using stats as part of the input, part of thinking about complexity and kind of having a way to explore complexity. And that's also been in my mind in my own life. So I'm excited to not just have a kind of these two different parallel comps, but how can they actually support each other? Right, right.
00:18:12
Speaker
So I had to come here for a week and cancel a week of class. I had to come to Paris. It was tough. The sacrifices you made. I know. I do legitimately take careful consideration if I'm going to not be around campus because I want to do right by my students.
00:18:29
Speaker
When I think about my takeaways, I think of what I'm bringing back. What am I bringing back? And I'm inspired by how much care is put into developing graphics, particularly when they're focused on issues like racism or climate change. And I'm really heartened that I'm coming back and showing awesome examples
00:18:48
Speaker
visualization applied to trying to understand and communicate about social issues rather than coming back and just showing like really cool whiz bang animations of whatever transition cool stuff which is awesome right and it drives engagement and it gets people excited but to be able to channel that excitement towards something that is valuable is the whole reason that I think many of us are here. This is a tiny thing after such an inspiring thing.
00:19:14
Speaker
that Mike just said. But one tiny thing is Sean Carter's not the only one to do this, but he did it really well in this talk of just using something to connect words to other stuff, each color in that case of like, oh, there's this highlight color and this equation, and now they're connected. And as a super specific thing, I'm going to do that a lot more. Yeah, that's great.
Trends in Data Visualization
00:19:37
Speaker
So I think the one last sort of general question because Mike you mentioned like the whiz bang and obviously we saw a lot of interactivity but it seems to me lately that there's been sort of a pullback in a lot of places from making the big whiz bang interactive thing and I even feel like here having been to this conference for multiple years that there seem to be even kind of a
00:20:00
Speaker
Not as big of emphasis on the big interactive that there was more like we made this map right like Erin Williams talked about segregation America the map itself it wasn't interactive and sort of the Since we might have had two or three years ago where you'd click on all the dots and something would pop up Do you see that do you feel that that people are starting to pull back a little bit from interactivity or are they using it differently?
00:20:23
Speaker
Yeah, definitely. I feel that I think people are realizing now that it's not necessarily the interactivity, it's the story that you want to tell that is most important and that maybe as journalists in general you have to guide people through that story just as tell them
00:20:42
Speaker
them on the focus that you want to keep the focus on exactly so that's maybe a reason why interactivity is less also because it is complex to do and that's why you see more of these storytelling or storytelling or just text with a visual and then some text and another visual doesn't have to all be in one visual
00:21:00
Speaker
although I have to admit it, I still very much appreciate, for example, the talk by Brian Jacobs from National Geographic about the amazing piece he did for the Cassini mission and the Dragon dinosaur thing. I'm still happy those things happen, but the thing is though, those two things are amazing, but they are also not that interactive. You again follow the story that they want to tell you, but I would say that the visuals themselves were really kind of a whiz-bang.
00:21:30
Speaker
Yeah, but not a lot of stuff to actually do with them. Yeah, that's kind of the direction I see more things going.
00:21:43
Speaker
We all know to butcher some Ben Scheiderman visualizations about answers to the pictures. We're not here to make more fun pictures. We're here to try and understand the picture. I think that perhaps we have invested, we the community have invested time in making things like cool and fun and we're realizing that it's not always helpful. So we are trying to find the best way to answer questions and sometimes that's interactive. But perhaps are the polish factor that has worn off and we're able to, you know, rain it in a little bit.
00:22:15
Speaker
as it can answer our questions. It's the truly live podcast. We're really holding down that, finishing the party. So yeah, I think also, I think it's kind of like we've had more experience over a much longer period of time of thinking intentionally about form and visualization. I always love historic fizz and thinking about what you can do. Like many, many hundreds of years of pen and pencil in this,
00:22:46
Speaker
More recently what we can do is we have computers and I think as interactivity came on the scene it was really exciting and it was kind of like we have this new sandbox and I think now people are starting to refine how we play in that sandbox and be able to kind of be more directive or more intentional about the types of motion or animation or how we manage time.
00:23:08
Speaker
in those experiences. So I think it's just kind of a really exciting time that people kind of explore these more intentional ways of engaging. Awesome.
00:23:17
Speaker
Terrific. Well, it sounds like everybody had a good time.
Conclusion and Gratitude
00:23:21
Speaker
So the last thing probably to mention for those of you who weren't here, Lynn Turney was the sort of program chair along with many others and also the Leon Business School who helped sponsor the conference. An amazing event for two days. They took, I think, good care of us. Yes. Good French food. A great lineup of speakers.
00:23:47
Speaker
So thanks to the three of you for coming on the show. Thanks to everyone for tuning into this week's episode. If you have comments or questions, please let me know. So until next time, this has been the policy of this podcast. Thanks so much for listening.