Julia Wolfe, Reuters: " It's helped remove a lot of cruft from my team's life"

Introduction

[00:00:04] Julia Wolfe: I'm super excited to talk about this. Because Datawrapper has done this for me, which I promise it will do for you: it's helped remove a lot of cruft from my life, from my team's life, so that we can focus on cool, original, bespoke things. I think, yeah, I'm just so grateful it exists. Thank you all. 

So, as Elliot mentioned, I am the Americas Graphics Editor at Reuters, which means I run a team of graphics journalists that covers news from the top of Canada down to the tip of Argentina. So, a pretty broad direction. But before I was that, before Datawrapper, as Elliot mentioned, he and I go way, way back.

So here is us, trying to, after a long cross-country journey, to get to SourceCon, a news conference, where you can see me and Elliot hanging out. I had more embarrassing photos, Elliot. You should be happy I went with this one. And our friend Andrew and somebody else, we're going to get back to later. He will show up again. 

My experience with Datawrapper

[00:01:11] Julia Wolfe: So a little bit about how I came to be a Datawrapper user advocate, acolyte. I was first introduced to Datawrapper when I was teaching at the Graduate Journalism School at CUNY, where it was used to teach the students how to make charts for the first time. And it was so exciting. I had seen charting tools before, but it really did democratize charting. These journalism students who were inundated with all these new tools, and they were really able to dive in. And we could put aside technical things, and really just focus on what it meant to make a good data visualization.

And so I was pretty excited about that. And then I went to work at 538 where we had an internal charting tool, but when I became the head of the data visualization team there, one of the first things I pushed through was retiring that tool and switching to Datawrapper. Talk about this a little more later, but that internal tool is just, it's just not worth it.

And I think it's pretty easily the biggest legacy I'll have left on a newsroom, switching that tool over. It made a world of a difference. And then I came to Reuters, where they were already on Datawrapper. But because it's an enormous newsroom, I am still onboarding people constantly. I support our newsroom, think of new ways to make the tool work for us.

So what I wanted to talk about today was my experience onboarding at 538, when it was a small newsroom and it was totally new and at Reuters, where it's an enormous global newsroom. Because the challenges are a little different at those two places, but hopefully I have some advice if you're already on that journey or you're thinking about taking that step, some tips from my experience of doing that.

Tips for introducing Datawrapper to a small newsroom

[00:02:52] Julia Wolfe: So I'll start with the small newsroom experience first, when you're introducing a small newsroom to Datawrapper or just a newsroom for the very first time that's never used it before. 

Build consensus

[00:03:03] Julia Wolfe: The first thing I did was really work to build consensus early. I felt pretty strongly we were switching to Datawrapper one way or another, but I didn't tell the whole newsroom that.

I got some of the reporters who liked helping us with charts. I got the copy editors to play with the free tool and let them believe that they had a lot of say with whether or not we went ahead with it. Because I trusted correctly that once they spent time with the tool, they would be really excited and they would back me up, in that this was something we had to do. So by the time the tool came around, everybody felt like they had been a part of bringing that on board. Which meant that, with any new tool, there's going to be some bumps. But people just really viewed those through good, optimistic, excited lens. And if, the people who hadn't maybe been brought on early, they didn't just have the graphics team advocating for it, they had people from different directions saying no, this is great for us. 

Use a project management system

[00:04:00] Julia Wolfe: The other thing I would suggest, which I didn't do right away, and I wish I had, would be to commit to a project management system. If you're going to be bringing this on for the first time, the Datawrapper folks are so wonderfully flexible, they'll work in whatever thing you want, they're not going to make you submit Jira tickets or whatever, if you've worked with different groups before, which is great in theory.

But there is so many. little details that go into making sure that the Datawrapper build for your newsroom has everything your newsroom needs. At 538, we had a very specific style we wanted to meet, which required a lot of back and forth tweaking. And I wish from the very beginning I had set up a really rigorous Notion board or GitHub project board, or whatever your tool of choice is. That just would have made it easier to keep track of that stuff. We did eventually switch over there and I think the Datawrapper folks were happy we did. That would be one tip. You are the one starting from the ground up. 

Think about locking features down

[00:05:04] Julia Wolfe: The other thing Datawrapper will offer you is the ability to lock some features down. And this is really going to vary newsroom to newsroom. But you should think very thoughtfully about what chart types you want, any particular use cases of those. Something that we deal with a lot at Reuters is reporters using the small multiple pie charts to make over time pie charts. So we've talked about maybe just removing that charting ability completely. If you think that there's some chart types that your newsroom might make, and you don't want to have to look over their shoulder at every single chart. You can remove features.

Show others how to use Datawrapper

[00:05:54] Julia Wolfe: The last is that in these early days, and especially in a small newsroom where you can do this, as many quick, let me show you that, let me hop on a call with you, as you can with different people who are using the tool. I highly recommend. I've brought so many people on board because it's a lot more exciting to watch me do something and feel like they can do it too, then to try to bump around in a tool or not know what's possible or not have those ideas. And so those early days, anytime somebody would ask me for help, I wouldn't just do it for them. I would say, let me show you. And we'd do it. They either be in person, pre COVID, or hop on a call, because that really got them excited and also let them start fixing things for themselves in the future.

Accept Datawrapper's help 

[00:06:39] Julia Wolfe: Finally, accept Datawrapper's wonderful help. They have fantastic newsroom training, which they ran for us at 538 and brought a lot of people on board. One thing that Elana had said to me at the time, and really stuck with me when I asked her for advice on this was that a lot of people think they have to be tier one customer support for their newsroom, but you don't have to be. Datawrapper will do that for you. Obviously there's the fantastic Datawrapper Academy documentation, but also in a small newsroom, if you're on Slack, they'll make a collaborative Slack channel. If I couldn't help a reporter, I was busy, I would say, just post in the Datawrapper Slack and somebody would get back to them. And so realizing that I didn't have to carry that weight was huge.

That was my quick lightning talk of the small newsroom. I'm trying to move quickly cause I know we don't have a ton of time and I want to make sure that if anybody is going through this now, they have time for questions.

Tips for using Datawrapper in a big organization

[00:07:42] Julia Wolfe: So the other half of this experience for me has been rolling onboarding at Reuters, which is a completely different challenge than 538. The tool's already there. We're on the enterprise level. But Reuters is a enormous global newsroom with people constantly joining or people who have been there for 20 years hearing about Datawrapper for the first time. Because you can do all the promotion you want, but in a newsroom that big and that spread out, lots of people won't even know it exists. And so I've learned a lot in the few years I've been at Reuters that have made things a lot better and made our charts or Datawrapper usage better.

Internal docs & email templates

[00:08:25] Julia Wolfe: Internal docs and email templates for the big newsroom is so important. Datawrapper has fantastic documentation, but it's not designed for your newsroom. We, for example, are a wire service with a whole other CMS just for our wire. We have rigged Datawrapper to publish directly to the wire. We have all sorts of technical things that aren't going to show up in Datawrapper's documentation and it would be a nightmare to keep walking people through that stuff all the time. And finally, it took me too long to do this, but I finally created an email template because the way that we add new users is they send to our listserv and I was writing the same text over and over to say, Hey, I've added you. Here is our link, and you can see on the left what that goes to, of our own documentation, including a link to Datawrapper documentation themselves. And also, in a small newsroom, I didn't have to worry about my colleagues remembering that I'm a real person, and not an AI chatbot or something. But in a large newsroom, sometimes I try to think about ways to add a little bit more humanity to the Datawrapper stuff, so they know there's actual people using this tool. They can ask for help. They can get excited about it.

Invest in your acolytes

[00:09:49] Julia Wolfe: The opposite of the small newsroom: you can't have quick, let me just show you, for every single person because there's tons of people in the newsroom and that would be all your time. So what I try to do instead is I identify people in different parts of the newsroom who are doing cool things with Datawrapper, are really interested in pushing how they use it, and I spend a lot of time with them. And those people have started spending a lot of time with the people on their teams. I had a desk editor who's been at Reuters for 20 years, no technical experience, who's one of these people for me, and he did a map of the fire in Texas, where he found the GeoJSON, downloaded it, uploaded it to Datawrapper, the kind of thing that I would have thought five years ago only the graphics team would get to that level of complexity, but he and I had spent so much time with the tool, and I assured him, that's something you can do. You don't have to be able to open that file. Just look for that geojson and the map ended up turning out fantastic. And he now helps the desk debug and handle and all that stuff, which has been great. As an example, I pulled all of our users and I sorted them by how many charts they've made. Number one is Ben Welsh. He and John are talking tomorrow, I think, about automated charts. So if you wonder why he's at 21k and everybody else is in the hundreds, thousands. You should go to his talk. But I have two people in here in our top five, top six, who are not even on the graphics team. But they make so many charts, they use Datawrapper just for their own exploratory reporting, but also for publishing. And that tells you how much this chart tool has meant for us. Because those are charts that, pre Datawrapper, they would have been asking our team to help with.

Tag team with production

[00:11:41] Julia Wolfe: Find your people who are in production. When you're running trainings and in a big newsroom, you're going to run a lot of trainings. And whenever possible, bring somebody from production with you, whoever it is. Desk, copy editor, a specific production team, whoever actually put stories together, they're going to know better than you what the common pitfalls are. Because a lot of graphics teams, the Reuters team included, we don't actually use the Reuter CMS and all that. We have our own server. So I don't know that much about how to debug when there's issues embedding a Datawrapper chart there. The production team does. And also, graphics teams seem very mysterious and technical to some folks in the giant newsroom. So they don't necessarily believe us when we say it's an easy to use tool. But when they hear that directly from the mouth of a desk editor or copy editor, someone they trust, saying, no, this tool is really easy, you can figure this out, I found they're a lot more likely to be less terrified and give it a go. 

Automate your charts

[00:12:49] Julia Wolfe: One more plug for Ben and John's talk, Datawrapper's API. It has taken so much work off of our plate, automating Datawrapper charts. Because we had to do the jobs report, earnings reports, FED stuff. That's just my region. But every region has their own cyclical, financial, whatever charts that are happening constantly. And since we've started automating those, that's just a lot of work that nobody in the newsroom has to do. We can tell reporters, don't worry. We actually had somebody from the legal team ask for some help on making jobs charts, job state charts for the legal sector, and I was like, oh let me connect you to Ben who already automates a bunch of those charts, and he can just add one for legal, you'll never have to make this chart again. He was pretty excited. Okay. And that's how you can bring Datawrapper to your small newsroom or to your big newsroom. 

Why use Datawrapper in a newsroom?

[00:13:54] Julia Wolfe: Very quickly, why should you bring Datawrapper to your newsroom? Your graphics team is magic. They have skills that nobody else has. They can code, they can draw, they can map, they can write, they can report, they can do 18 different things. Do not have them spending their time making bar charts. Because the thing that I believe sets the great graphics teams apart from the good graphics teams is largely determined by how much cruft you keep off their plates. And so if you can get those daily charts away from them, if you can get maintenance of stuff away from them, if you can turn more reporters to the tool, more automation of the tool, that means your team is doing more original, bespoke graphics journalism that no one else is doing. And that is a much better use of their time and your team.

Internal tools are too much work! Elliot and I can tell you all about the joys of the journal's internal charting tool, and the maintenance of that, and asking for help on maintaining that. At 538, it was the same thing. Before, I was at Reuters they had a tool and it's basically become somebody's full time job to support. But probably somebody who does not have a background in software, who does not have the setup that Datawrapper has, you're not going to end up saving money when you end up needing employees spending full time maintaining a tool. Or not maintaining the tool, which begins to fall apart and have a million bugs, and not do what you need.

So yeah, don't have an internal tool. Use Datawrapper. It's great. 

Collaboration

[00:15:39] Julia Wolfe: One last thing. It's great for collaboration. That person who I said would come back, this was a very important email. We sent the Datawrapper team about a project we were working on together. Which was our child, because, the wonderful people at Reuters sent me and him this great onesie. That, you'll have to get the file from somebody at Reuters. I don't know if the Datawrapper folks have it. and it was really cute and sweet. And that's it. That was 15 minutes, supposed to be 20, but I said, maybe I could leave time for more questions. So yes, I would love anybody has been working on this or has any questions about what it's like to onboard Datawrapper into your newsroom, big or small, or teach with it. I'd love to talk about that because I love Datawrapper. Thank you Datawrapper people.

[00:16:38] Elliot Bentley: Thanks so much, Julia. That Baby wrapper is one of the best things that's ever landed in my inbox, I have to say..

Q: Maintaining quality when non-graphics people make graphics?

[00:16:46] Julia Wolfe: When you allow non graphics people to make graphics, how do I maintain quality? This is such a good question. When I was at 538, we were a small newsroom, so we reviewed every single chart. And it still went through a chart edit and that was possible at 538 because we published five stories a day and not every story had a chart in it. 

That's not possible at Reuters. We have taken a bit of a calculated risk in that we don't have a formal review system other than we can look at all the Datawrapper charts. What we do is nothing automatically moves to clients. The graphics team has to approve to move it to clients. But it does mean that charts can end up on reuters.com that we are not wild about. And that gets me to my point about locking some things down. I think in retrospect, maybe we should go back to the Datawrapper folks and turn off some charts that I see some problem.

Q: Reduced quality of visualizations?

[00:17:42] Elliot Bentley: I wanted to jump in with a question of my own, which is, what would you say to someone worried that putting charting tools in the hands of reporters might result in a kind of reduced quality of charts?

[00:17:59] Julia Wolfe: It does. Yeah, like that's the trade off, right? I think, if I look at the charts made by my team... If we didn't let the newsroom make charts, and only we made charts, the charts on Reuters would be better. There would be 1/50th of them. The reality is that we are a small team in a large newsroom, which is true I think for most graphics teams relative to the size of their newsroom.

And I think it's a choice as an editor. You're always deciding where to spend your time. And I just really strongly believe that your graphics team's time is better spent on the high value work only they can do and let some mediocre charts go through. Like when we can help out, we will, I really encourage people always to reach out if they're not sure if something's working. Luckily, we've trained up enough of the desk to help us with any of the big misleading problematic charts. But yeah, do I see some spaghetti charts with 30 colors sometimes? Sure. Do I see some things we would have done differently? Yes. But to brag a little, the Reuters graphics team was named by two different orgs last year the best team in the world.

And I think part of how we reach that milestone is that we just don't spend time on things that other people can do for us. 

Q: Supervision for non-graphics people?

[00:19:17] Elliot Bentley: Let's see, Laura asks, When someone from outside the graphics team makes a chart, is there any supervision? And do you set boundaries on which types of charts the non- graphics members can do? For example, only tables and bar charts, or do they have freedom to create?

[00:19:34] Julia Wolfe: Yeah, there is not any official supervision other than the fact that it'll go through an edit from their editor and from the copy desk, like the story itself. So there is in that sense, but there isn't necessarily chart supervision. Again, this is different at 538. If you're in a small newsroom and you can look at every single chart, great, but we couldn't do that at Reuters.

We don't currently lock down anything, at least on the news side. I think Datawrapper might have some charts for other kinds of clients, but if I was going to, honestly, the only one I would probably look down is the small multiple pie charts, because that's the only one I've seen. And I'm not like a pie chart hater, but I've seen some wild uses of that one, and I'm not sure the newsroom needs that one. So if we're going to lock down, I think that would be the one.

Q: Other tools in use?

[00:20:25] Elliot Bentley: Okay, so another question from Patricia asks, What other tools did you use in your team?

[00:20:34] Julia Wolfe: What tools do we use? We don't use a ton of... We use Blender, it's a free 3D modeling tool, for some of our 3D. Again, that's specifically the graphics team. The newsroom is not using Blender. I use the programming language R, so I use R studio to make charts sometimes. Although I will say Datawrapper's small multiple line feature has made R suddenly less important for my workflow because I think that was the moment I would switch out of Datawrapper and into R. We use Illustrator, of course. I mean with Datawrapper, you can export to SVG. So often a workflow for us will be using Datawrapper to get the data in and then bring it into Illustrator to touch it up or do something special.

Otherwise we're doing a lot of custom coding in JavaScript, with Svelte, which is what our rig is built on. Which is stuff that we have time for because of Datawrapper. 

Q: Getting non-graphics folks to use Datawrapper?

[00:21:33] Elliot Bentley: Great, alright, here's a classic question from Connie. Do you have tips about getting non graphics folks to use the tool? There's sometimes a gap between who I train and who actually makes charts.

[00:21:45] Julia Wolfe: Yes. Yeah. And again, that really varies to my mind on the size of your newsroom. I think if it's the small newsroom, get on the call with as many people as you possibly can. And if people who haven't reached out to you much, just say, I'd love to see how you're using it. 

Tougher on the big newsroom. So there, you can see in Datawrapper, if you're saying who I'm training and who's using it, you can see in Datawrapper who's using it a lot. So you can reach out to those folks proactively. I think that whole like be human thing I shared, that's been really important, because Reuters is huge. They don't know who's friendly, who's helpful. And I try to be very super friendly whenever I do those things and remind them, I'm here if you want to, if you have any questions, I'll do a lot of demos where somebody's struggling that I think I can help them with and make them watch me turn a meh chart into a great one only using Datawrapper and there's something so soothing about watching someone who's good at something do it.

So that often gets them very excited about Datawrapper, and makes them realize. And it just gives me a chance to teach a lot of Data Vis 101 stuff. But the other thing is, don't worry. Some people just won't really use it, and that's okay, I think that's the other thing. It's great that we have it, but I'm not out there trying to make every reporter put Datawrapper charts in their stories. I'm just trying to make the ones that want charts in their stories bug us less.

Q: Template system?

[00:23:08] Elliot Bentley: I've got a next question is from Natasja, I hope I said that correctly. Hi Julia, fantastic speech. Do you use some kind of template system for people outside of the graphics team?

[00:23:23] Julia Wolfe: No. I mean, we have, but I think when you bring Datawrapper into your newsroom on the paid accounts, they will style the charts. So we do have a template in that sense that all of our charts come in the Reuters fonts, we have some default colors for them to choose from, we have the way that we format our dateline and notes. And all that is in our template.

And that was the same thing for 538. That's why I like that PMing process was so important because getting the styles just right was a lot of back and forth. But now the template is just when they go to make a new chart and Datawrapper, it comes with all that stuff.

Q: Resistance to being told to use Datawrapper?

[00:24:04] Elliot Bentley: Great. We have time for a few more questions. I see one from Gregor from Datawrapper, who says, I'm curious if you experienced any cases where people in the newsroom showed forms of resistance to being forced to use a new tool, being introduced in a sort of top down way?

[00:24:27] Julia Wolfe: That's why in the small newsroom, I did that before Datawrapper was officially brought on board, I made several people, including some people who I knew would love it, and some people who I expected that resistance from, brought them on before we had signed up Datawrapper, so that they felt like they had the chance to discuss concerns. And honestly, at the risk of sounding like a shill, it's such a good tool that once people use it, the resistance tends to go away. And because there's a free version of the tool, you can get people using it right away. At Reuters, I think I just go back to that voluntary aspect when people seem stressed. I know the reporters are really under a lot of stress, and so if they feel like, I don't have time for this right now, I say, that's fine. We also don't have time to make a chart for you, so then your story just won't have a chart in it, but I am not somebody that believes every news story has to have a chart in it. So I think when they realize nobody's going to force it down their throat, it's just another option for them, I can see the like resistance melt away and then they might give it a shot when they're less busy.

[00:25:36] Elliot Bentley: Every news story needing a chart. I've never heard that idea before in my life.

[00:25:41] Julia Wolfe: Ugh. You have only worked with better editors than me, Elliot. There are some people outside of graphics teams who think apparently that's a good goal.

Q: Style guide for non-graphics people?

[00:25:51] Elliot Bentley: Okay, so a question from Will. Do you give the non- graphics people a data visualization style guide, or do you communicate best practices in other ways?

[00:26:01] Julia Wolfe:  So the style guide we don't have to do as much because the base styles are built into Datawrapper, thankfully. And the bounds of Datawrapper keep them from going too wild. But we do have in our documentation a Dataviz 101 talk that we give a lot. So we have different Datawrapper trainings that we give, but whenever we give any Datawrapper training, we always make some time at least to just... you know, the big things we want people to know about charting. Like some tips on color, some tips on when to use certain chart forms, when to not, the things that we see a lot that we're more concerned about journalistically. And honestly, I send a lot of people to Lisa's writing on the Datawrapper blog, because that does a lot of the work for me. So she has some great write ups about really picking your highlight colors cleverly and what a line chart is good for. We'll give those talks, but then additionally I'll often send them a very specific article if they have follow up questions about that from Lisa.

Q: How fast are journalist when news break?

[00:27:04] Elliot Bentley: Cool. Another question here. We're almost at the end of our slot, but lots of great questions coming in. Richard asks, how quickly do journalists publish their charts after a non-scheduled news event?

[00:27:17] Julia Wolfe: This is great: So fast, if they want to, because we have connected Datawrapper directly to our wire, to our CMS. So when they hit the Datawrapper publish button, it goes automatically into our internal CMS. So at that point, the desk can themselves move it to the wire. There's also an option that just the graphics team can use from Datawrapper to even skip that and immediately push it live to the wire. So as quickly as they can make a chart in Datawrapper, which is very quickly, it can be on reuters.com. It can be to our clients. It's lovely. 

[00:27:59] Elliot Bentley: Okay. I think that is all the questions in the chat. Thanks so much, Julia, for coming in. 

[00:28:07] Julia Wolfe: Thank you. Good to see you again, Elliot!

[00:28:09] Elliot Bentley: Yes, you too.