Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Episode #58: The Financial Times Graphic Desk
182 Plays8 years ago
Welcome back to the show! This week, I’m very pleased to be joined by John Burn-Murdoch, Alan Smith, and Martin Stabe from the graphics desk at the Financial Times. Established in 1884, the Financial Times is one of the world’s leading...
The post Episode #58: The Financial Times Graphic Desk appeared first on PolicyViz.
Recommended
Transcript
Introduction to PolicyViz Podcast
00:00:00
Speaker
This episode of the PolicyViz podcast is brought to you by Tableau Software. Tableau helps people see and understand their data. Tableau 10 is the latest version of the company's rapid-fire, easy-to-use visual analytics software. It includes a completely refreshed design, mobile enhancements, new options for preparing, integrating, and connecting to data, and a host of new enterprise capabilities. To learn more, visit tableau.com.
Meet the FT Data Visualization Experts
00:00:39
Speaker
Welcome back to the Policy Vis podcast. I'm your host, John Schwabisch. I'm joined by a trio of like-minded data and data vis enthusiasts this week from the Financial Times, the graphic desk over there. I'm joined by Alan Smith, John Byrne Murdock, and Martin Staba, and we'll get to Martin's last name momentarily. We'll come back to that. Guys, thanks for coming on the show.
00:01:00
Speaker
Pleasure to be here. Thanks for having us. I'm looking forward to talking with you guys about a lot of different things across the newsroom and how the graphics desk works with the newsroom and the workflow that you have going
Alan Smith's Transition to FT
00:01:12
Speaker
on there. But before we dive into that, maybe I could have each of you sort of introduce yourselves, give a quick recap of your careers and how you ended up at FT. Alan, maybe we can start with you.
00:01:23
Speaker
Sure. I'm the data visualization editor at the EFT. I've been here almost exactly a year. My path to the newsroom is unusual because I spent most of my career in the civil service at the UK's Office for National Statistics. I'm the EFT's first data vis editor. My role at the moment is
00:01:43
Speaker
I've been particularly in the first year looking at how we bring together print and online and what were previously quite disparate data viz enterprises into more of a single function. I'm Martin.
Martin Staba on Workflow Integration
00:01:57
Speaker
I'm the head of the interaction news team at the FT. I've been here for about six years now.
00:02:02
Speaker
What that means is kind of changing, you know, as, as our workflows are coming together between the sort of web focused teams that I've run for the last couple of years and what Alan's been doing more on the print side recently and bringing those to get to together. But historically what it meant was we were kind of the web focused side of the graphics operation and the more sort of code driven side of the graphics operation.
John Byrne Murdock on Team Collaboration
00:02:31
Speaker
And I'm John. I think the last check of my Twitter bio, I'm a database journalist. And before that was interactive data journalist. So some kind of mixture of that, I guess. So working physically on Martin's team, but I guess across the two disciplines in terms of
00:02:48
Speaker
telling stories through graphics, sometimes they're interactive, sometimes they're not. Right. So you all sort of mentioned this separation, or I guess traditional or historical separation between the print side and the media and the website, and now sort of trying to bring them together. So can you talk a little bit about what that workflow looks like? In addition, not just within and across your teams, but also with the reporters who are out in the field making calls, telling stories. Take us through the journey of a story as it goes through the entire process.
00:03:16
Speaker
I guess the really lame answer to that is it depends a little bit. It always depends, right?
Enhancing Visual Storytelling Through Collaboration
00:03:23
Speaker
It's true, though. Obviously, being based in the newsroom, we've got a relationship built with the different desks here in an editorial sense.
00:03:32
Speaker
The editors here, for example, on markets, companies, world news, UK news, you know, we're building relationships with those people where we are trying to, I suppose, think about graphics, not so much as like a kiosk in the corner of the newsroom that's dispensing graphics to order, but trying to work with those people, you know, maybe earlier in the development of a story. So we start to identify some opportunities for graphics as early as possible. And some stories will lend themselves more to that.
00:04:01
Speaker
than others essentially the kind of life cycle of our graphics is traditionally also kind of varied a little bit depending on where it was going so i kind of martin mentioned it briefly but i guess the kind of online graphics originally started as a slightly separate enterprise and because of that has been very very separate workflows depending on whether you want your graphic to appear just in print or as an online piece
Evolving Role of Print at FT
00:04:27
Speaker
I guess just kind of organizationally, the role of print is changing a little bit. So, you know, logically, it's rather than the place to get your breaking news, print is, if you like, the best of the last 24 hours online. So there's some sort of inherent motivators to bring those things closer together. So I mean, one of the things I've been trying to do on that side of things is to try and make sure that we're using a single production system, regardless you're generating, you know, charts for social media,
00:04:55
Speaker
interactive content or just simply graphics for print. So go back to what I said originally, it does depend. We work a lot with our editors, but we're also trying to make sure that we work a lot with our reporters to try and
00:05:07
Speaker
kind of instill in our reporters the idea, the trigger early on, that there's a data and or a visual component to a piece that needs working on as an integral part of the story's development.
Training Journalists in Data Visualization
00:05:19
Speaker
And that's much more effective than that. I've written my story, can you just drop a chart in at the end, which is what we're trying to kind of move away from. Right.
00:05:26
Speaker
And you've written in a recent chart doctor, we had talked about different chart types for that article. So how are you working with the journalists to get them to think more visually as opposed to, oh, I'm just going to whatever the easiest bar chart is? How do you get that data visualization sort of ethos ingrained in their brains? That's a great question. I mean, in fact, in some ways, that was been one of the things that since I arrived, I really wanted to get tucked into.
00:05:51
Speaker
One of the things we introduced here as part of our digital newsroom training agenda was a course we developed, just a very simple course on not so much how to make charts or to actually do data visualization, but we called it the competent critic, the idea that someone should be able to pick up a newspaper or read a website and actually look at a chart and decide, and to be able to deconstruct it, does it actually work or not? And can I articulate the reasons why it does or doesn't work?
00:06:20
Speaker
So we put together this course to try and do that. And the way I wanted it to work was to try and instill in our journalists that.
00:06:29
Speaker
these graphics were depicting relationships in data that were core to our stories. And that's where kind of our paths crossed John on the graphic continuum, because I kind of decided that we needed something very, very similar to the graphic continuum to sort of help articulate that message across the newsroom.
Core Skills for Data Visualization Roles
00:06:44
Speaker
And writing it in the Chart Doctor series is quite good, because as well as sharing it with FT readers, it's a nice way of then publishing it back to the newsroom, if you like, so everyone gets to see it.
00:06:55
Speaker
So from your perspective, either print or web, when you are looking for people, either in the newsroom or on the graphics desk, what are the core skill sets you're looking for? In terms of hiring, I guess? Yeah, I mean, I think traditionally, there were kind of two camps, right? We had a graphics team that hired from graphic design backgrounds,
00:07:20
Speaker
And we had on the interactive team, we had kind of the sort of what might have in the past been called computer assisted reporters, people with a journalism background, but a sort of data analytical bent.
00:07:34
Speaker
and developers, so people who are writing front-end code to display graphics on the web, but also designers who are very focused on the UX elements and this kind of triad of skills of design
00:07:52
Speaker
development and data analysis was kind of the three sectors that we brought people together from. I think the idea that that isn't needed on the print side is kind of going away. I heard one of your previous episodes, you had Kennedy Elliott from the Long Washington Post talking about very similarly how a lot of the tools and techniques that we've long used
00:08:15
Speaker
to create interactive graphics for the web being just as useful to just simply create geometry for print graphics. So a lot of the structures that we've long established to make graphics for the web are now also being used to make graphics that are static and that are not interactive or animated in any way.
Unified Production System for Graphics
00:08:36
Speaker
So those three gillsets are now valuable regardless of what medium you're primarily working for.
00:08:43
Speaker
And just to complete that as well, I think what's interesting is that the tools that had been used reflected the fact that people were coming from different backgrounds. So for example, with a team who were working largely with print, their chart creation tools were effectively like a 25-year-old Illustrator plug-in with a line bar or buy. And as Martin says, if people kind of wanted something a little bit different, then they tended to associate that with interactive
00:09:12
Speaker
content that was being designed from scratch. And so what we've really wanted to do is make sure that the toolkit encourages people to come closer together. So the fact that we're using sort of D3 now to drive print as well as online graphics kind of again helps us to think about the skill sets in a much more integrated way, which certainly wasn't traditionally the case.
00:09:36
Speaker
But what I find interesting, especially John, some of your work that I just love is how you take what might be sort of a natural place to add a whole bunch of interactivity and you make it static and you add just very simple annotation on top of it.
00:09:55
Speaker
top
The Debate: Static vs Interactive Graphics
00:09:56
Speaker
of my head is one you did on the London Marathon that splits between men and women and you had, you know, column for men and column for women. He had changes over time and it was a scatterplot. And so those sorts of things when you have these scatterplots, they just sort of like the first thing I think a lot of people think about is I'm going to add a layer of interactivity to this so that everyone can click on every dot. And I'm curious because
00:10:15
Speaker
I see a lot of the things that you guys produce is static and it's not sort of relying on the interactivity. So what is your view now between the balance between static representations and ones that sort of require or don't require interactivity?
00:10:28
Speaker
Yeah, I mean, speaking purely on those kind of examples that I've been involved in, I think some of it is absolutely a fairly clear idea about the direction we want to go in on this. But I'll hold my hands up as well and say that some of it is often convenience. So things like the fact that the core CMS that the FT uses for web publishing doesn't allow for interactivity. So there will be times where
00:10:53
Speaker
essentially out of feeling guilty for building my own bespoke web pages outside the CMS, I'll think, you know, this one should be something that fits in our existing system. And therefore, I'm limited to static graphics. But at the same time, there's absolutely a sort of whole ideal behind that as well, which is very much in line with what people like Archie say, at the New York Times have said in terms of
00:11:20
Speaker
I think the growing consensus that interactivity is often a little bit gratuitous and you can make huge time and resource savings in from not including that step but also in terms of the load it might put on your web browser or the very fact that having to click on or hover over a dot to get some information is essentially asking the reader to jump through a hoop that they shouldn't have to do. It should be our
Interactive Graphics in the Newsroom
00:11:44
Speaker
job as journalists to pick out
00:11:47
Speaker
the key parts of a given visualization and surface that information straight away. And I, my view is sort of that the, and again, many people have had this line of thinking before is that the annotation layer is where the journalism really comes into to visual journalism. It's sort of making a graphic is the equivalent to sort of interviewing your source, but it's then your job to actually pick out the bits just as a writer would pick out the key quotes, the key
00:12:15
Speaker
arguments, you have got to pick out the key bits of this data set that the reader should know about. And the only case when I think I've had to sort of check myself on that, as it were, is when I'll publish a scatterplot as a classic example. And I will have picked out what I think to the average reader are going to be the key points that they need to be aware of. But there's always going to be one reader will then go to something and say, Oh, I would have been great if this was interactive, because I really wanted to know what that dot was or this dot was. So I think
00:12:45
Speaker
It's important still to keep an open mind and think about one of the cases that people are going to want to ignore this from their own individual perspective. But yeah, generally, I think we're moving in the same direction as the industry as a whole in terms of thinking that there are essentially quite a lot of reasons that people were using interaction in the past, which weren't catering for readers needs, and were more sort of doing it because they could. One of the things that we've talked
00:13:09
Speaker
about even before the vocabulary concept came in, was trying to teach people under what circumstances to use interactivity. And we kind of boiled that down to a set of rules that we tried to teach again in courses that we ran about how to commission graphics effectively.
00:13:25
Speaker
One of the things we realized is that for a long time, interactivity was almost used as a status symbol. It was used as a signal to readers that we thought this was important and worth throwing a lot of resources at. You got a lot of gratuitous interactivity as a result of that. Weaning the newsroom off of that has been one of our long-term education processes.
00:13:46
Speaker
One of the things we're saying now and as part of the competent critic stuff is, you know, interactive is an adjective, not a noun. In the past, a lot of times people would say, you know, I'd like to build an interactive. Well, no, actually, you want a graphic that has the attribute of interactivity. And perhaps you don't even mean interactivity in the sense of
00:14:09
Speaker
click here to progress through a story, or, you know, some other thing that requires some custom interface design, you might just mean something that is automated, or something that is dynamic in some other way or personalized, again, in some automated way, maybe rather than having to search for your postcode to find a localized version of the story, we might use your phone's
00:14:33
Speaker
location awareness to predetermine what you're most likely to be interested in based on your current location. So some of the things that in the past might have been interactive are now personalized or are dynamic or are automated. We're trying to refine the vocabulary we use to describe basically graphics that move in some way or that require some sort of reader feedback rather than something that's purely authored by us and read in a sort of linear way.
00:15:03
Speaker
So kind of reading people off the idea of interactive as a thing that is somehow distinct from a graphic that just happens to be static.
Aligning Print and Online Graphics
00:15:12
Speaker
That's really interesting. So how does it, when you think about both static and a graph that has interactivity on it, so trying to use your vocabulary, that appears on the on the FT website versus what appears in the print version, is that a whole other layer of separation or is that your whole goal is to try to interleave all those together so that they're all in the same discussion?
00:15:35
Speaker
The plan is very much so behind our visual vocabulary. There's a lot of developing repo of chart content that is scaffolded for the different platforms. It allows us to build for the different platforms at the same time. If you know you're going to be working on something that's going to require some online and print treatment, and they're going to be different, you can treat them as such at the time you're making it. You don't have the separate workflows for it, and the two can develop in parallel.
00:16:02
Speaker
Which is much better than the sort of sequential work that we've had before you know so much closer together now in terms of dna and what we don't want to do is homogenize it cuz actually one of the things is great listening to nick susan is a tapestry celebrating large format print you know for things like comics and so on and it's the same here with the ft if you're printing.
00:16:24
Speaker
a quarter of a million broadsheet copies a day, you want to be able to use that space in a way that... The content should be different, but as much as possible, where it's got a data-driven workflow, you want it to be efficient too. That's still developing, but that's something I'm quite passionate about, is that we build in efficiencies where we can in terms of the production, but we don't shortcut the design as a result.
00:16:48
Speaker
I want to turn back a minute because, John, you had mentioned sometimes you put out a complex graph and you have the annotation. There's always one or two people, sort of the outliers who, you know, I want to know this data point. I want this value. And I'm curious what the stance is or what the strategy is at FT about openness in terms of the data, in terms of the design.
00:17:10
Speaker
You're obviously open when it comes to the visual vocabulary, which you've published and allowing people to download and use and edit.
FT's Data Openness and Sharing
00:17:16
Speaker
But I'm curious to what extent you are, when you can, where you can post data and code. So what's the openness of the newsroom in terms of getting things out to the broader public?
00:17:28
Speaker
Another really good question, and I think it very much depends on the data set. So being the FT, a lot of the stuff we work with on a daily basis comes from providers that we pay generally on a subscription basis to get that data from.
00:17:44
Speaker
In those cases, it's not really our place to put that data, or in some cases, it's contractually absolutely not our place to publish that data. The graphic is essentially the best we can do. There are others where we've compiled data sets.
00:18:01
Speaker
through our own research and that data was initially public. The best example I can think of this was a piece that our colleague Robin Kwong did a couple of months ago on spending on private jets from board members and CEOs of major companies. He spent a very long time putting this data set together, structuring what was previously spread across hundreds of documents.
00:18:27
Speaker
Because that data was public initially and because we felt, or he felt, there may be value in putting it out there for readers to find their own points of interest in, that was published wholesale. And I think that's generally what we feel, speaking from my own perspective at least, is that if the data was public and could still provide a public use, we'll put it out there. If it's not, there's less of a justification for it. I wouldn't say we sort of have a negative stance on this thing. I think a lot of the time it's just
00:18:55
Speaker
the way things work. We finish a project and we're on to the next one. On my own work, there are definitely examples where I could have published data sets and haven't. That's generally been purely through the time available to me. Although I will also hold my hand up and say that there are definitely cases where I will intentionally not publish
00:19:18
Speaker
a data set that I've spent a week compiling sort of on some of the sports stuff I'm doing. If I think that there's a good chance that I'll get additional stories from that data set over time, I will selfishly not publish that for exactly that reason.
00:19:34
Speaker
I mean, I'll definitely say that that's in the minority of times, but I do think this is something that often isn't discussed. The fact that if you spent a long time putting some together, invested your time and therefore your employer's money into compiling a large amount of information, there are cases where I think it should be justified for you to say, we're going to hold this back for now.
00:19:57
Speaker
Once you think you've got it out of it, I think, yeah, you should put it out there. 538, obviously, are fantastic at putting all their information, all their data sets up on their GitHub page. I think that's the ideal, but there are cases where I can see why people don't publish. Right. I mean, if you're spending a lot of time pulling together dozens of public data sets, that's a lot of, as you said, that's a lot of time and money that you've put in. So that now has value that you would want to try to take advantage of.
00:20:23
Speaker
Yeah. Am I right in saying that ProPublica have a tiered system where most of the data sets they make available for free and then some of them, if non-public data was used or if a certain amount of time was invested in putting the data set together, they'll charge a nominal amount, I think it's $10 or something. That was certainly the case at one point. I think that's an interesting approach as well, saying this is free, but in a case where we've invested time and resources, we are going to ask for a bit of money.
00:20:51
Speaker
Yeah, right. Great.
Public Engagement with Data Visualizations
00:20:53
Speaker
Well, I want to kind of totally switch gears here because we've been talking about what's going on in the newsroom and the evolution of data journalism at FT. But we've spoken in the past about people's perceptions of policy and data and whether people have more or less faith in data, whether they have sort of the knowledge and the skill sets required to fully grasp and understand some of this sort of
00:21:16
Speaker
more complicated topics or more complicated graph types. And so I'm curious whether you have thoughts on that coming from a newsroom where your readership is probably more of a targeted readership. Maybe they certainly have more expertise in certain areas than in others. What are you seeing when you get feedback from your readers in terms of what they bring to the table and what they're walking away with? What's their level of sophistication?
00:21:43
Speaker
Where do we as sort of a society need to take data visualization and data and statistics education so that people sort of can grasp and understand these complex topics? I know that's a huge question, and I know we're not going to solve it, but I'm just curious what you're seeing from sort of your readers and even from people in the newsroom. I think if I can, I mean, it is a big question, this, but I mean, I'm going to step back a little bit as well away from just looking at FT readers to start with, because I think
00:22:12
Speaker
We're having this conversation at a time where, in fact, in some ways it's been a really pretty bad time for data in a public sort of sense. I mean, certainly here in the UK with the recent EU referendum.
00:22:23
Speaker
We saw some really pretty catastrophic uses and perceptions of data.
Challenges in Public Data Understanding
00:22:28
Speaker
It was quite disappointing to see the way that data was treated by the people involved in the very public debates. In fact, I think probably the most memorable figure from the whole debate was a fictitious one or it was a gross figure advertised as a net figure. Very basic numeracy that still got shouted about a lot and actually seemed to have some big pull on the debate.
00:22:49
Speaker
I mean, I think where I was before in the statistics office, it was really interesting because I used to get very frustrated that people knew a lot about data and I knew a lot about what the data was saying because they'd compiled it were very often prevented from kind of logically continuing that understanding of data into the public arena. And I thought that firstly was a shame, you know, that taxpayer funded expertise very often didn't go the final sort of six centimeters in terms of getting information back into people's heads.
00:23:19
Speaker
Here at the FT, it's really interesting. I think the reaction, for example, to the Chart Doctor series has just shown you what a range of different types of personalities you have, even among the FT readership. Some people will see something that we've produced and ask for something, a bit like John was saying, in terms of access to underlying data and interactives, for example, they'll want more.
00:23:43
Speaker
And whereas some people will say, well, actually, you know, that chord diagrams way too complicated, I want something much simpler. So it's very difficult to sort of abstract it down to just a single sort of persona or a single sort of person that you're writing for. I think
00:24:00
Speaker
There's a slight paradox in that having said that it hasn't been a very good time for data. I mean, from the FT's perspective, one of the things that really surprised me was when we wrote the the Chart Doctor article, the most recent one, and I talked to our social media team about what was our most engaging tweet of the year. It was one of John Scatterplots, you know, which actually was seen as like, well, this is more complicated to interpret, but because it was topical, because it was well annotated and because it had the text to sort of explain the context.
00:24:29
Speaker
There you go, that's your most popular tweet out of 24,000 tweets that we've published this year. That was good evidence that with the right context, with the right explanation, and certainly with the right sort of topicality,
00:24:42
Speaker
we can achieve great reach and engagement with our readers. I don't know if you wanted to say anything more about that, John, on that. That particular one, I think your spot on in terms of it was the setting and the context and the text as well of that that made it resonate. I think it was an emotive issue. It was a scatter plot pointing out that there was an interesting pattern whereby regions who exported a lot of their economic output
00:25:09
Speaker
to the European Union happened to be the ones with the highest votes to leave the EU. And there were obviously 48.1% of Britain was pretty in a bit of a shock and surprise and the fact that they could turn to this and say, what have you done? It made it made it work really well. Yeah, so I think it wasn't necessarily the fact that it was a scatterplot. But being a scatterplot and doing as well as it did, I think was a really pleasant surprise for us.
00:25:38
Speaker
I think there's certain types of visualizations that traditionally have raised eyebrows among even within our own newsroom. I'm thinking of things, scatter plots are definitely one, chord diagrams, even cartograms of various sorts that are sort of non geographical in their description of a geographical data set have traditionally often raised questions of isn't this too complicated for our readers to understand quickly.
00:26:05
Speaker
And one of the things that Alan said in the first chart doctor column was given a strong enough dataset and a rich enough story that's been appropriately visualized.
Fostering Deep Engagement with Visuals
00:26:17
Speaker
It's completely appropriate to expect someone to spend a couple of minutes learning how to use a chart and reading it, right? And just as you would expect someone to invest a couple of minutes of time in a long form.
00:26:30
Speaker
piece of written content, a really richly produced graphic can be something that someone lingers over rather than just glances at in the in the passing context of a bundle where the text is the primary communications media. Just picking up on that, there's a daily feature in the FT, the big read, which was where I did deliberately targeted the first court diagram that we were going to do because I said, look, we're calling this the big read, we're actually saying sit down and read this, you know, grab a couple
00:26:59
Speaker
that you're a beer or whatever you need and spend some time looking at this. In fact, in that first Chart Doctor article that Martin was referring to, it was great to pick up on people like Scott Klein and Alberto Cairo and all these people saying the same thing, which is don't try and dumb down what you're doing so that everything has to be glanceable. Sometimes you need people to read things. In some ways, there's no greater feedback for me than the person who says,
00:27:25
Speaker
I didn't get that first, but then I spent time looking at it and then it made sense. And then you think, well, I couldn't have articulated it into any other form that you could have understood in two or three seconds. So that was a worthwhile thing, both for us as the producer and for you as the reader to do. I think that's tremendously important because if we don't have that attitude agreed across the newsroom, it will drastically reduce our capabilities to sort of have a lasting
Conclusion and Audience Engagement
00:27:49
Speaker
impact on FT content. We need to get past that barrier. Yeah, absolutely. That is really interesting. Jens, this has been fascinating. Thanks so much for coming on the show. Great. Thanks a lot, John. Thanks, everyone, for tuning in to this week's episode. If you have comments or suggestions, please let me know on Twitter or on the website. Of course, please
00:28:11
Speaker
do review the show on iTunes, Stitcher, or your favorite podcast provider. So until next week, this has been the PolicyViz Podcast. Thanks so much for listening.
00:28:31
Speaker
This episode of the PolicyViz podcast is brought to you by Tableau Software. Tableau helps people see and understand their data. Tableau 10 is the latest version of the company's rapid fire, easy to use visual analytics software. It includes a completely refreshed design, mobile enhancements, new options for preparing, integrating, and connecting to data, and a host of new enterprise capabilities. To learn more, visit tableau.com.