Part of the FT’s interactive news team, John-Burn Murdoch works as a journalist alongside developers and designers to produce a mix of long term data-driven projects and same day interactive news stories. Other activities include presenting to domestic and overseas conferences on data journalism, data protection and big data.

JBM’s recent “bar chart race” caused quite the stir in the data visualization community and it was a real treat for me to spend a bit of time talking to him about his process to create this and other visualizations. We talk about his bar chart race and animation in data visualization more generally.

Episode Notes

JBM | Original Bar Chart Race Tweet | Narrated version on FT

Tweets on the Bar Chart Race | Ollie | Ajit Niranjan | Dan Cookson | Neil Richards

The Data Debate: Andy Kirk v. Andy Cotgreave

Spanish Grand Prix, Josh Katz at the New York Times

Tools | Flourish Bar Chart Race tool | gganimate | Tableau

Papers:

Effectiveness of Animation in Trend Visualization by Robertson et al

Animated Transitions in Statistical Data Graphics by Heer and Robertson

Enhancing Information Visualization with Motion by Bartram

“Effectiveness of Animation in Trend Visualization,” ten years later from Danyel Fisher

If you’re in Washington, DC on Wednesday, June 18, 2019, I’ll be talking about animation in data visualization to kickoff the SAGE Ocean series here in DC.

Support the Show

I recently revised the steps in my Patreon page, so please check it out and see if you’d like to become a financial supporter of the show. If even for a few months, your support helps me cover audio editing services, transcription services, and more. You can also support the show by sharing it with others and reviewing it on iTunes or your favorite podcast provider.

Transcript

Welcome back to the PolicyViz podcast. I am your host, Jon Schwabish. On this week’s episode, we are going to talk about animation and data visualization and the advantages and disadvantages of taking data visualizations and animating them as things transition from one state to another. To help me talk about that, I’ve invited John Burn-Murdoch from the Financial Times to join me on the show. John recently created a Bar Chart Race as it’s become fondly called, that showed changes in populations of different cities as they change over time, and that Bar Chart Race got a lot of attention. So John and I sat down and talked about that project that he created, why he created it, and all the other ways in which animation can be really beneficial to presenting and communicating data. In fact, I am about to give a talk on animating data visualization tomorrow night, here in Washington DC, as part of the SAGE Ocean Speaker Series, and I’m really excited to be a part of it to help kick off the Washington DC series. Hopefully, later this fall, I will be in London to give a talk at the London office of SAGE, so I’m looking forward to that as well.

So before I get into the show, just a couple of quick notes. If you’re interested in supporting the show, please do share it, tell your friends, tell your family. If you’d like to financially support the show to help me cover transcription costs and audio editing costs, please check out my Patreon page. I have a number of different new platforms that you can use to support the show, all the way from $1 a month to $3 a month, $5, $10, and even more. If you’re so willing, I’d really appreciate it. If you’re interested in learning more about data visualization, check out my workshops page. Later this fall, I will be conducting a bunch of public workshops. In September, I’ll be teaming up with Stefanie Posavec now one of the members of the Dear Data team, and we will be conducting our Dear Data series, we’re going to take it on a road show, we’re going to start in New York I think, we’re going to head over to Philly, and then we’re going to end up here in DC. In October, I’m teaming up with a friend of mine, here in DC, to data visualization and R workshop. Finally, later in the fall, I will be in Amsterdam to teach a data visualization and excel workshop for a full day with my friends over there at Graphic Hunters. If you’re interested in attending any of those sessions and workshops, please do check out the public workshops page on my website. If you have any questions or you want me to come and give a workshop to your organization, please feel free to reach out and let me know.

Back to the podcast, on this week’s episode, I’m talking with John Burn-Murdoch from the Financial Times about animating data and here’s that interview.

Jon Schwabish: Hey JBM. How are you doing?

John Burn-Murdoch: I’m good. How are you?

JS: I’m good. I like this JBM moniker, I can call you. I feel like…

JBM: I’ve been getting [inaudible 00:03:15] between the two Johns I guess.

JS: Yeah. It does a little bit, it does. How are you? You just got back for vacation?

JBM: I did, yeah. I was over in Italy for a few days, which was fantastic. It was a real joy to be out there, so yeah.

JS: Lovely. Stefanie Posavec and I have done these workshops and she’s originally from the United States, from Colorado, and she’s in London, and we were going to meet in Chicago for something like a year ago, and I said, oh yeah, she was going to be there, her husband’s family’s out there, and I said, there’s no problem, I’ll just fly there, it’s like two and a half hours flight, no big deal. She said two and half hours, that’s like from London to Turkey. So when you’re like, I just went to Italy for whatever, that’s like me going to Delaware. [inaudible 00:03:57]

JBM: Yeah.

JS: I want to chat with you on the show because you hit, I think, a nerve with your Bar Chart Races, and people really responded well to them, and then Flourish came out with their own ability to really drop and drag and make something about them. So I was hoping maybe you could talk a little bit about why you built it and that process, and then we can talk about animation sort of more generally and your thoughts on the value of animating DataViz.

JBM: Sure. Awesome. Yeah, so this was an interesting one. The one that I put out there which I think seemed to get a lot of traction was – I think it was March 18 or 19 that I put that one out – and that was because in the couple of weeks leading up to that, I’d just been seeing, as I think a lot of people both within and outside the dates of his community had been seeing, a lot of these animated bar charts or Bar Chart Races, whatever we want to call them, doing the rounds on social media. I think one of the first ones that really seems to have sort of come out of nowhere was this one showing the estimated values of the biggest brands in the world and how they’ve changed over the last 20 or so years. I spoke to Matthew Navarro who’d shared that one on Twitter, I spoke to him earlier today, and we’re trying to work out where that one had originated, and he wasn’t exactly sure either, so he just picked it up off YouTube or something and then put it out. So there’d been that one, and then there was one showing the number of goals that different soccer players had scored when they were all at a certain age, and these just seemed to be coming out of nowhere, and coming out of different corners of the internet, different software, different stylistic touches, but the format just seemed to be really catching fire. There was huge, huge traction around these, and it just got me thinking. I guess, I was thinking about it from two different perspectives. One was just the simple like, oh I want to know can I make one of these in the various toolkits that I’m aware of. The second was more of a fundamental like, is there really something about this format which makes these graphics really resonate, and so I wanted to sort of find that out for myself.

So step one for me was to take the global brand values one and actually simply recreate that one in D3, so using an observable, like the [inaudible 00:06:31] contouring sharing platform because I already had the data for that one; I thought, well, this would just be a proof of concept, see if I can get the old D3 transitions up and running again and make sure it all works. I did that fairly quickly and thought, okay, this works; and then I thought, all right, what would be another interesting dataset to use in this format to create something of my own. Initially, I was thinking I’ll do sort of the size of different national economies and what have been the top 10 economies over time, so I was working with that data; but yeah, that seemed fine, that seemed interesting, but then I stumbled across this incredible dataset of populations of pretty much every city that has been on earth going back thousands of years. I just thought, so you’ve got – first of all, the data just fits very nicely, it’s a dataset where you’ve got lots of different entities and each of them has a value that changes over time so that ticks a very basic requirement for what I guess we will keep calling a Bar Chart Race. The magnitudes of these numbers change a lot, so you go from populations of thousands to hundreds of thousands to tens of millions. I guess, the other thing I really liked about this, I felt it was just an inherently sort of curiosity peaking dataset. Everyone’s fascinated by which the largest cities in the world are, I think plenty of people probably couldn’t even tell you the largest city in the world today. But certainly, when you throw that back 500-600 years, it gets just really, really fascinating. So as soon as I started digging into the data, I thought, yeah, this is going to be the one.

I then plugged that dataset into the same observable template that I built for the original brands one, and I added a few additional bells and whistles. I think that was when I put on a world map which sort of served as a color legend for the collars of the bars, I did each continent in a different color, and allowed me to sort of have these little pings on the map whenever a new city entered the top 10. So that was pretty much that, and I did the whole of that, so from sort of coming up with the idea to making the brands example and the city example, that was essentially just playing around over a weekend. And then I think, I set them loose, as it were, on Twitter, on the Monday morning, and I posted it to Reddit as well, and then pretty much just sat back and my phone proceeded to melt as all the notifications came in. So that was essentially it. It was a little weekend side project and, I guess for me, it really demonstrated that when you’ve got the right data to use in that format, these things, you know, people just go absolutely crazy for them.

JS: You and I had a quick conversation about it after it came out, and I was thinking, well, you could just do a bump chart with this, but then when you have these – like you said, you have a lot of data over a lot of years, you kind of get this spaghetti looking thing, but I’m curious what should we think about this. You have to sit and watch this thing for two minutes, it’s not the immediate takeaway that you would get from a line chart or a bar chart, and I’m curious what you think, it strikes people that we’re in a fast content world and yet they’re willing to spend two, two-and-a-half minutes watching these bar charts move around and just grow and move up and down. Why are people willing to spend two, two-and-a-half minutes watching these things?

JBM: Yeah, it’s really interesting, and that probably strikes as the most counterintuitive thing, because we’re in a world today, and especially on social media, where attention span is about four seconds. So the idea that people would sit back and watch is crazy. So I’ve jotted down a few of the responses it got, and one person said exactly this, they said, I can’t put my finger on why these are so fascinating, you’ve got to watch for far longer than it would take to look at a line graph, so why is this so engaging. But then someone responded to them saying that the fact that you’ve got this time to speculate on the changes that you’re seeing is for them the best part of this whole format. So the graph itself, the chart itself isn’t answering the whys but it’s allowing your mind to think, oh wait, so this thing just went up, that thing went down, I wonder what was happening there. The fact that it essentially keeps you sitting in front of it thinking about what you’re seeing, that is essentially a luxury in the DataViz world. We all want to think that someone is sitting down with our charts for at least a few seconds to think about what they’re showing. But very often, with a static chart, someone will have a quick look, and if the overall message comes up clearly, which is a good thing, then they may well just say, okay, got it, and move on; whereas, with a chart like this, the fact that you have to sit there and watch, it does seem and makes sense to me, but it also seems from people saying this without being prompted that that act of sitting there watching does give you that time to really immerse yourself in the subject matter and think about what you’re seeing which you might not otherwise get. And then someone else said, they wonder if the format encourages the analytical and questioning behaviors in a viewer that a scientist or data analyst might do instinctively with a static graph. But maybe within an audience of people who aren’t familiar with the idea of reading a chart and consuming it as a piece of information, with that audience, the animation essentially is what holds them captive and gets them thinking more inquisitively. So I think both of those are really interesting points about why the fact that you’re spending time with it ends up being a strength rather than a weakness.

Separately, and just there’s this broader question of should these not simply be line charts all the time, so Andy Cotgreave of Tableau fame among the other hat he wears, he and Andy Kirk had a great little debate about this at a London Tableau meetup a month or so back where they were each just given once – they either had to be pro or against Bar Chart Races and Andy Kirk

[inaudible 00:12:59]

and Andy Cotgreave, he used a great line I think actually, which is, he said Bar Chart Races are like the fidgets spinners of DataViz. These are things that everyone, for a while, goes crazy about and spends a lot of time with, and then they end up being a fad. It’s an interesting argument to be made there, and only time will tell as to whether these just end up being a passing fad. But there are a few things that came up in that debate which I think are interesting points, that sort of got me thinking more about what it is that makes these work and cases when they do work well and cases when they don’t work as well. I think I’ve kind of distilled it to three sort of ingredients, as it were, in the [inaudible 00:13:39] and the story you want to tell that are really well suited to this format. If one or more of those ingredients are in place, then this could be the perfect way to tell the story; and if they’re not, then maybe it should just be a line chart. So one thing I think is you should really be using these when the values in your dataset are changing hugely over time, when you’ve got orders of magnitude and an exponential growth. So a lot of people reacted to the one that I first put out showing city populations and said, yeah, why not just make this a line chart; and they wanted to be able to trace, for example, the population of Cairo or of New Yorker of London over time, rather than having it on screen one second and not the next. But the problem you have gone into immediately with that is the populations of these cities are just vastly, vastly different from one carrier to the next.

So when the time series starts in the year 1500, the largest city in the world has a population of 672,000, which is today 38 million. If you had a standard line chart going from 0 to 38 million, then in 1500, that 600,000, you’re not going to be able to see it at all, and you’re certainly not going to be able to differentiate between the largest city on 672,000 and the second on 500,000, straightaway you have an issue there. Some people then counter and say, okay, but what if you had a logarithmic scale. First of all, you’d have to have that logarithmic scale changing over time, but even then the tiny differences between a city with 200,000 and 185,000 population, again, that really come through. So for me, one of the huge strengths of the Bar Chart Race, where you are essentially saying that your horizontal numeric axis is constantly clamped at the highest value in the dataset, the huge advantage is that you get to view every year whether it’s 1500 or 2018, you get to view every snapshot of your data as if that’s all that matters, as if you were alive at that time, and that’s what you’re looking at. So it allows you to constantly view this data, to essentially see the most important points of the data as they were at every different point in time, instead of having to view everything from the perspective of what’s important today which is…

JS: Tens of millions, yeah.

JBM: Right. So one thing I think when you’ve got dataset, when you’ve got a time series that changes by several orders of magnitude, I think this is perfect. Another thing, and this is one of a couple of points where I’m probably going to make analogies to movies and TV, I think these work really well when you’ve got a changing cast of characters when certain things drop in and out of those rankings over time. So again, with a line chart, the problem you might have is that in order to show every city that has been at a certain point in time in the 10 most popular cities in the world, you would have to have a 100 odd lines on that chart. So that again is not really particularly readable. Whereas, if you want to be able to show 10 cities in 1500, but then a different 10 cities in 1600, a different 10 in 1800, etc., this format again works really well. So if your points of interest or, as I’m saying it, your characters are sort of going on and off stage and changing, then animation is really the only way to do that while still having a sort of accessible number of data points on there.

The third thing that I think works really well is when your audience really can get on board with the idea quite naturally, get on board with the idea that they’re watching some kind of race unfold. So anything where it’s ranking, so like the original one that piqued my interest which was the most valuable brands in the world; or, in this case, the largest city in the world, if you can give someone the idea that who is at the top matters in some way, then I think that just gives that – you get a little bit of emotional investment, the idea that you’re watching a race. I just feel like people are so conditioned to watching races unfold whether it’s watching the Olympics every four years or whatever, that we just – this is a format that we just are familiar with, and so we get that tension that suspense; and this is another thing that several people commented on was that you really do feel a sense of excitement because you don’t know what’s coming around the corner in all of those things together. I realize, I mean, a lot of what I’ve been talking about here is more about the pure sort of emotional reaction that someone has to these rather than I’m not talking purely about perception and the usual things we talk about when evaluating DataViz; and I’m sure, we’ll come on to that. But just in terms of people seeing this and just being completely immersed in it for the two minutes or so it takes to watch, I think when you have those things, so changing orders of magnitude, changing data points of interest, and this idea that you’re watching some kind of race unfold, I really do think this is the best way of showing that data. I don’t think that that is something which is only true now because it’s a new thing we’re seeing, I think that’s just a fundamental – just a fact when you’re displaying this kind of data.

JS: I totally agree with all of that and I think there’s, on this point of getting behind it, there’s this other part of, similar to a map, at least the one that you did, at least the one on populations, this holds for a lot of the other ones, it also allows you like a map to place yourself in the context of the data. For example, when I’m watching it, growing up in America, I’m looking for – you see New York show up, and then you say, oh that’s when the country started, you can see it grow, you can see different cities in America, you can see San Francisco show up and then sort of fade off; and, like you said, you can imagine those stories and you can identify with your particular area of the world, I’m sure other people in other continents and countries have the same experience of watching, understanding the history of their own country and seeing the cities rise and fall of where they live and where they are. So I think there’s that personal connection with these sorts of things that works well too.

JBM: Yeah, 100%, and I guess that’s important to talk here about how, yes, this particular visual format seems to perform really well, but I think a lot of that is about what’s being shown. So the original one showing brands, people could look at it and think, oh IBM, yeah, I forgot, they were so big; and then, oh, look at Amazon that we have now; people can recognize the entities there and it’s exactly the same here. I think the fact that this was looking at global cities meant that people from all over the world who, even those who can’t necessarily read English, they’re still probably going to recognize the English spelling of the city they come from, or cities in the country. So I’ve seen people have taken this video and made their own Facebook posts in more than a dozen languages that I’ve seen from various Indian scripts to Japanese and Chinese. I think, the dual things of number one long time series, so people get that curiosity that you get when you’re discovering about the past; and number two, all of these recognizable places on the planet, I think those were really, really helpful as sort of hooks for people to get into the story, as it were. I think if this were the same format and even the same numerical values, but looking at something like the most, I don’t know, the most prescribed medical product over time or something outside of the pharmaceutical industry, I suspect, people might be a lot less engaged with it. So the subject matter, I think, definitely is part of the reason some of these have been really popular. But I think when you put it all together with this format as well, then this version of this data is probably just going to be more engaging than a static one.

JS: Right. I sort of came to know and love your work because of what I view as sort of an expert way of using annotation on your graphs. But a lot of those are static, and so I wonder how you and maybe the rest of the staff there feel about or approach animation sort of more generally. We’ve been talking sort of specifically about the Bar Chart Race, but the lots of types of animation obviously, and so I just wonder what your experience is and what your thoughts are on DataViz animation which is just a totally different – I mean, maybe not a totally different thing, but it’s certainly a different way to approach communicating data to people.

JBM: It’s a good question because again, just staying with the city’s example for a second, I think one of the strengths of it almost is the fact that there isn’t a commentary over it and that allows individual people to focus on whichever part of it resonates most with them. So we did, at the FT, we did do a second cut of this with a voiceover from me on there as well, and I’ve done a couple of live performances of it, as it were, with commentaries. I think those have been great for other reasons, it adds even more of this sort of sense of suspense and excitement when you’ve got a commentary over the top. But I think what you lose there in this case is people are then probably going to be focusing on whatever I or another narrator is highlighting; whereas with no commentary, they’re free to look at whatever bit they think is interesting to them at a given time. So that’s just on this specific example. More generally, I and my colleagues here at the FT, we really do think one of the most valuable things we can do as data visualization practitioners is add this expert annotation layer. And so with these kind of animations when you’re just seeing sort of a data story unfold, as it were, I think voiceovers are absolutely one way to go.

I know that Flourish who, as we’ve discussed, have put out their own template for this as well. I know they have, in their toolkit, you can add voiceovers to animations. You can do that using text on the screen as well instead of obviously doing it through audio. So a subsequent animated piece that we did here, maybe about a month after this one, was looking at the time of the day that different people around Europe eat meals; and for that one, I added a bit to the observable notebook that I’ve created so that I could pause the animation and put up some short text annotations on-screen which I think helped. More broadly, that’s something that we’ve been doing in our animations for the last couple of years, so several of my colleagues here will often do animated pieces where you’re essentially taking one static chart or map and just showing a couple of key frames which really sort of highlight the story that is being told there. In some cases, they will be showing something that did change over time. So we might have a map showing, for example, some military activity in Syria; and frame one will show how things were in, say, January; and then frame two will highlight an event happening in February. In other cases, we’re taking one chart that doesn’t actually change but just putting up sequential annotations to guide someone through the chart.

I guess, the two things we tend to focus on there is, number one, make sure the amount of text that is in any given annotation on an animation is short. So you’re not asking someone to sort of stare into the screen and read a couple of paragraphs, but you’re just putting, say, a maximum of a dozen or so words up that they’re going to look at for a second or two, and also just to make sure you get the pacing right. So I don’t think we have a hard and fast rule on this. But in one of these animations, so take the map example again, where we’re going to be showing, say, four or five events and putting up a sentence on each one showing what happened, well, pointing to a certain part of the map, I think we generally go for about a couple of seconds’ pause when each of those annotations comes up; because the difficulty really is as soon as you have a, you transition away from an annotation before someone’s ready, that’s just such a frustrating user experience; and in all likelihood, someone’s not going to bother with sort of dragging the video slider back, they’re just going to click away, especially when some of these are GIFs rather than videos and you don’t even have the ability to scroll…

JS: You don’t have any controls to use it.

JBM: Right. So yeah, I think it’s with animation, you’ve really got a design for someone who isn’t going to scrub across the video, even though they can, you’ve got to design it for someone who just wants to sit back and make sure that you get the timings right so that they really can take in everything without feeling that they’ve been interrupted essentially.

JS: So what do you think it is about the animation that makes it popular? Is it that it’s more passive receiving the information in sort of a more passive way, you don’t actually have to examine the chart in the same way because you’re adding this annotation, you’re telling the story, is it because it’s like a movie or video which people might sort of instinctively associate with watching a movie or watching a television show like, what is it do you think about the animation that gets people to stick around for longer than they might with the same graph or even a series of graphs that are that are more static?

JBM: Sure. So I think there’s bits of all of that and a few other things in here as well. First of all, just in terms of the initial task of getting the audience’s attention, as we’ve discussed at the beginning, people are famously short of attention these days, and that’s no more the case that it is on social media which is where a lot of data visualization is consumed. I think the fact that movement is a very, very effective way of drawing someone’s eye, it’s probably a big part of just getting the initial eyeballs on the graphic, and that’s step one to producing an effective piece of data visualization. You can produce, as Hans Rosling said in an interview with the Financial Times a few years ago, you can produce the best piece of work ever but if people don’t actually pay any attention then, as a visual communicator, you’ve failed. I think the fact that movement, that motion really draws people’s eyes is the first part. One of the papers I was looking at, when I wanted to dig some more into this, it’s a paper called Enhancing Visualizations with Motion, and that is by, let me just see if I’ve still got that on screen, but if not, anyway moving on – yeah, it’s from a paper called Enhancing Visualizations with Motion and the researcher found that when comparing color, different colors, different shapes or using a little bit of motion to attract a data visualization viewer’s attention to a certain data point, using motion, even very small motions or slow motions was far more effective at drawing attention, and this remained true when they [inaudible 00:29:55] read back the values that were being conveyed by the points that were moving.

So first of all, I think there’s just that that basic fact that motion draws our eyes. It’s the same reason that the horrible websites that we all hate will spin up loads of auto-playing video ads. I’ve realized I’m in danger there saying that Bar Chart Races are sort of horrible data [inaudible 00:30:17]. But yes, so on the very basic level, I think, motion draws attention. But then another paper, a really interesting paper that I read was one called the Effectiveness of Animation in Trend Visualization, and this is a paper by George Robertson and some colleagues that actually came out in 2008. But in 2018 IEEE VIS won an award for papers that really have had lasting value on the field; and they took the example of the now famous and much-loved Gapminder animated scatterplots Rosalind and Co. produced a few years ago now. They presented three different versions of that same dataset to participants and asked them several questions about how they felt about the effectiveness and enjoyment of the visualizations as well as their ability to read off the values. They had one animated version, one static version but using snail trails like the connected scatterplot, and one small multiples version; and they found that there is just something fun essentially about the animated version. I’m just reading through, they came back to it 10 years on and wrote up some of the most memorable insights they got from it, and so there’s a line here where they say that study participants describe the animated version as fun, exciting, and even emotionally touching, but at the same time some participants did find it confusing. They also found that whether or not the different displays of information had interactivity, users who had delivered the animated version were less accurate in their readings in the data than those that saw the traces of the small multiples.

There’s a few interesting things in that. So first of all, from this DataViz purist perspective of are people able to accurately read values of this visualization, it seems fairly clear that animated versions are actually slightly worse at that. I think that that seems pretty intuitive, when things are moving around, it’s harder to get a measure on them. But if you’ve also got people coming away and saying they really enjoyed the experience of consuming a visualization, they found it fun, exciting, emotionally touching, and they gave it more focus, then I think that’s noteworthy as well. When you look at the actual numeric values that they assigned to people reading off the values on the different charts as well, it was higher for the static examples but not vastly so. I think there’s a lot there that sort of tells us why people really do just seem to engage with frames that have motion in them. So coming back to what you talked about, is there something here about the familiarity of watching something unfold, do we think it’s like a movie or like a TV show, I think – again, it’s not just what I think but based on a lot of the responses to the piece that I put out and then a few others have done as well, the suspense of seeing something unfold really seems to hold someone’s attention far better than static charts do. The way I think about it is a line chart, yes, a line chart is going to be a more efficient way of communicating that information because everything is there upon the first viewing. But the flip side of that is that one way of talking about it would be to say that the Bar Chart Race is the no spoilers version of a line chart; and it’s that very fact that you don’t know what’s coming next, which again especially for people who aren’t familiar with spending a lot of time reading DataViz, that really just naturally draws people in and holds people in, in a way that a static chart doesn’t. I guess, my way of thinking about it is, it’s not so much that that one thing is better than the other, but I think recognizing and appreciating that animation, there is just something about animation which makes people want to stay with a data visualization, actually those who are not DataViz familiar, should we say, I think that’s really valuable, really worth knowing. And something that a couple of the academic papers I’ve read, talk about maybe it’s not about saying you should do one instead of the other, it’s about thinking about the situation you use them in. So maybe you say, well, on social media, we’re going to use the Bar Chart Race version because that’s where we’re probably going to encounter a lay audience, well, that is less familiar with DataViz and that we just want to engage and hold and inform about this. But then the version that I might use in a slide presentation when I’ve already got a captive audience or in a news story or an academic paper where the person reading the paper is already there and you don’t have to grab their attention, then that’s maybe where you use the static version that they can then pour over, analyze, and read off specific individual numbers. But I do think it worth appreciating that there are fundamental things about animation that are simply more effective than static animations in achieving certain results. Again, that’s why I think Andy Cotgreave lines that these are a fad, I think that’s unlikely to prove true.

JS: I want to ask one last question before we go. I would guess there are a few people listening to this episode saying, this all sounds good, I really like the Bar Chart Race, I agree with JBM about making animations but I have no idea how to do it or I don’t necessarily have the platform to create something animated. Are there techniques or tools that you either use or know of that you would point those folks to, and I’m thinking of the researcher or the data scientist who’s either in R program or data program or just using Excel, but have some good data and they want to make something, they think a quick little animation would really help them?

JBM: Sure. One of the really nice things about this has been that I’ve seen these made in so many different tools, and the first two, the brand’s one and the soccer players’ one, I don’t even know what software was used to make those. But I’ve certainly seen a lot of people making them using the Gganimate library in R. I’ve seen even some making these in MATLAB. I’ve seen some made in After Effects. Again, I made mine using the D3 JavaScript, and there’s this Open Notebook on Observable that anyone can fork and iterate on and use their own data. So yeah, there’s a lot of tools out there to use; and I guess one other I should mention is, and I’ve seen people even managing to do this with the animations in Tableau, and I know as well that another sort of halfway house version of this or at least a version that would probably deliver a lot of the same benefits would be to just actually make a set of small multiples, and then using one of the various free online tools for converting a set of images into an animated GIF, you could go about it like that. So take a set of snapshots for each year, for example, in your data, and then turn it into a GIF. So you wouldn’t get the same sort of smooth transitions from year to year, but you might still achieve that overall effect of watching a dataset sort of grow over time. One other example I should just mention while we’re at it, partly in terms of other people and tools producing this stuff, but also, I guess I presented a bit of a false dichotomy earlier in saying that this is a case of Bar Chart Race versus line charts, because a really, really great example of someone’s rating on this was Josh Katz at the New York Times has been doing some line chart races, as it were, showing how the Formula One motor races have been unfolding this year. So it’s the same principle of having these lines extend from the left to the right of a graphic and [inaudible 00:39:04]. But yeah, so Josh has been doing the same thing using lines instead of bars. So you probably gain a few of the elements that people have criticized Bar Chart Races for not showing, and it gets a bit closer to what some of the line chart advocates have been asking for. So lots of tools, lots of formats, lots of people doing this kind of stuff. Before I glaringly forget, Flourish, of course, have come out and created their own right template explicitly for making Bar Chart Races, so there’s lots of tools out there for people who code, for people who don’t code. So it’d be great to see more people experimenting with this kind of thing.

JS: Yeah, that’s great. This is great. I’m excited to see what you and your team at the Times come up with in the animation space. It’s certainly an interesting space to see developments. So thanks for taking some time out to chat with me about it.

JBM: No problem. Cheers.

JS: Cheers.

Thanks everyone for tuning in to this week’s episode. I hope you enjoyed that. I hope you learned something about animating data visualizations. I hope you’ll check out some of the links on the show notes page including the Patreon page that I’ve set up to help support the show. So until next time, this has been the PolicyViz podcast. Thanks so much for listening.

June 18, 2019
Leave a comment
4678