Simon Rogers is an award-winning data journalist, writer and speaker. Author of ‘Facts are Sacred‘, published by Faber & Faber in the UK, China and South Korea. He has also written a range of infographics for children books from Candlewick. Data editor on the News Lab team at Google, based in San Francisco, he is director of the Data Journalism Awards and teaches Data Journalism at Medill-Northwestern University in San Francisco and has taught at U Cal Berkeley Journalism school.
+ Full Transcript
Rosemary Pennington: We swim in a virtual stew of data. From voting data, to health data, to crime data, there’s likely a number somewhere that, with a little crunching, can get you a better understanding of your community. The task of crunching those numbers and making them intelligible for a broad audience is the focus of this episode of Stats and Stories, where we discuss the statistics behind the stories and the stories behind the statistics. I’m Rosemary Pennington. Stats and Stories is a production of Miami University’s departments of Statistics and Media, Journalism and Film, as well as the American Statistical Association. Joining me on the studio are regular panelists John Bailer, Chair of Miami Statistics Department, and Richard Campbell, former Chair of Media, Journalism and Film. Our guest today is Simon Rogers. Rogers is a data journalist who authored the book Facts Are Sacred, which quote, “reveals how data has changed our world and what it tells us”. He’s currently a data editor at Google, and before joining that new media giant, Rogers served as a data editor at Twitter, and worked at The Guardian where he created the newspaper’s data blog. Simon, thank you so much for being here.
Simon Rogers: Thank you so much for having me.
Pennington: How did data become such a big part of your life?
Rogers: Well, if you’d seen me when I was like ten, you would not have thought that this is how it was going to work out.
[Laughter]
Rogers: Like I knew I wanted to be a journalist, right? Like that was not a challenge for me, right? Soon as I stopped wanting to be a fighter pilot, I wanted to be a journalist. Math was not my strong point, and there’s this example I use when I do my talks, but basically- I have a math school report which says tries hard but has little natural ability. [laughter] It was not really what I imagined happening. But now when I think about how I used to think about numbers and so on, it makes a lot more sense to me. And I guess I kind of grew up in a house where I was always asking questions, I wanted to know why things happened a lot. I loved reading Richard Scarry books when I was a kid. And if you look at the cut-away drawings of Richard Scarry books, they are incredible to me now. And even now when I look at them with my son, and they are amazingly accurate visualizations, just drawn in a very accessible, friendly way. And I think that kind of representing information in an accessible way is something that anybody who works in data should strive for. So, I guess I kind of grew up wondering why things were the way they were, but not seeing math or numbers as a way to do that. But fate has a way of intervening, doesn’t it? So for me, I had joined The Guardian to edit the new news website- it was just when we launched it, at the end of the 90s, and I managed to get transferred to the newspaper which was like the center of gravity then, you know, so that shows how old I am. But my first day was September the 10th, 2001. So day two, 12:15 our time, everybody was out at lunch and suddenly the terrible events of that day, September 11th unfold, and basically they needed somebody to work with the graphics team and nobody else cared or wanted to, and I thought well, it’s always good to take an opportunity that everybody else needs but nobody else knows how to do, so I started working with the graphics team a lot and I started just collecting data. Collecting data partly because I’m a bit lazy, and the first time you look for GDP figures you don’t understand what any of these numbers are, why different measures of GDP is ridiculous. And then you get the right set, or carbon emissions or whatever it is, and you have it. So, we started collecting all these data sets, and after a while I just had loads of data. And I thought well what if we just published, on a blog, published datasets. You know, kind of format it in a useful way. So non-PDFs, you know just in a way that people can access news. So that’s what we started doing and that’s how I got into this strange field, I guess, is just by luck. By luck and just you know doing stuff in the office that nobody else wanted to do.
John Bailer: Well, I have to tell you that anybody that can put on a dataset that’s every Dr. Who villain since 1963 is all right in my book.
Rogers: Yeah, you know when I left, I really- I just said to the team, whatever you do just keep updating that.
Bailer: That has to be current –
Rogers: I mean we’re talking about why-
Bailer: Yeah, I think that’s just brilliant. So, you said that you quickly started working with- you were one of the first people to be enlisted to work with some of the designers there. What was the first story that you worked on?
Rogers: The first one that I remember- and I actually came across this when we moved to the U.S. because I was sorting our staff, and the Guardian had this thing at the time where you could have a big visualization, a big picture across the spread, and it would be a whole thing. And there is still, I’m kind of old fashioned about this, there is something about seeing data on paper. There’s something amazingly powerful about that. So anyway, I was lucky enough to work with The Guardians head of graphic design called Michael Robertson, and who’s just a brilliant designer, and at the head of the invasion of Iraq- and we’d done a lot of small things around Afghanistan and what was going on there, and we had to do a big thing that would hold the spread. And really it was about pulling together the right data, and realizing there’s a whole skill set that people don’t teach enough in journalism school, which is how to work with designers
Richard Campbell: I agree.
Rogers: Right? Because what people tend to do, and I’ve seen people do it, is a journalist would just give a designer a whole sheet of paper or whatever it is – tons of datasets, they’d say “here you go!” and then say “goodbye”. And the thing about this job I got I guess was I was encouraged to be an editor working with a designer. So, the thing that I would be editing would be the content and see how stuff fits together. So, I really learned quickly not to do that, but you have to edit content before you give it to designers and talk through what the visual should do and what story it is trying to tell and so on. And I teach data journalism at the Medill-Northwestern School in San Francisco and we have a whole session on how you work with designers. I think it’s something that everybody who works with data should learn how to do.
Pennington: And I’m stealing that idea from you right now. I teach journalism so…
Campbell: So, Simon, John and I got into this ten years ago when we taught a class on news and numbers, and I can remember as the journalism professor in the classroom I was a little intimidated because I had math anxiety. And later I’m going to talk to you about overcoming that, and what you do with students who are like that, because a lot of journalism students are very much like that. But in terms of design, I remember I felt comfortable with John in the very first class because he brought a graphic, a visualization in the class, he looked at the students and pointed at the board and said, “what’s the story here?”, and I felt really comfortable that John expected that a good data visualization should tell a story and you should see it right away. And I saw your TED talk, and this is the one from six or seven years ago and you had that nice – I call it a bubble chart- is that what you call that? Yeah, the one where all the money is spent and in Great Britain. So, what is bad design? I mean when you teach, when you look at a graph or a chart what are you telling – how do you get the story out there? How do you – what do you say to journalists?
Rogers: You know, I’m not a designer. Right? I’m still not a designer. Even though I care about visuals, I guess, but for me anything that makes you feel stupid and that is inaccessible and – there are some chart types I don’t really like I don’t really like. I don’t really like scatter plots, I’m sorry there are sometimes-
Pennington: You should see John’s face right now I think you’ve just broken his heart.
Rogers: Oh, I’m sorry.
[Laughter]
Rogers: I know that- I’ve seen them done like on signs and they do an incredible job-
Campbell: I don’t like them either
[Crosstalk/laughter]
Rogers: I’m not a fundamentalist about this, but I think for me it’s really about accessibility. And I mean for example, I think there was like a phase, a few years ago when all a visual had to do to go viral was kind of be pretty.
Bailer: Yes.
Rogers: And I think we’re past that now, because I think that what people really want to understand what’s happening in the world and this crazy time we’re living in, and I think any visual that doesn’t try to speak to everyone, then it’s about are you making this visual because you want to impress your peers? Or are you making it because you want to express information in a way that people can understand? So, I’m obviously – I’ve got an example in mind but, I think that sometimes I look at a visual and I think, okay I’ve spent 20 minutes trying to decide what the hell this is about, and while I’m doing that, and I’m used to charts, like, then what’s the point?
Bailer: So, a lot of people talk about storytelling with data. So, can you-
Rogers: Storytelling is like one of the most overused words.
Bailer: Yeah, it seems like it’s really getting a lot of play these days. So you talk about a data story, you talk about the data journalism you defined as the process to get the story out there that’s in the numbers, I mean, help us think about through the way people are talking about it these days. Whether it’s data journalism or storytelling with data or what is a data story?
Rogers: So, a few years ago to be a data journalist was like a separate career path. And now increasingly I think we’re seeing a better, more widespread sense of the importance of using data in journalism among all kinds of reporters. It means like you don’t always have to be an expert, you don’t have to be a coder or designer to be able to tell a story with data. I asked somebody about this – I did a Twitter thing where I said hey Twitter, how would you describe data journalism more recently, and really what came out to Me is that a lot of it is literally about telling a story in the best possible way, because you are not limiting yourself to doing it just with words. So, what other information can we bring in to enhance the story, make it more understandable more accessible. And I’m really psyched to see that in the kinds of work that’s out there among people who are- some people who are telling stories around data and they’re incredibly kind of human way. Like you look at some of the work of Mona Chalabi, who is the U.S. data editor at The Guardian now. And so she- I worked with her back when I was there, because she came to a training and so I’d see her as she came into the office, and she’s developed this whole field now of drawing datasets in a way that are incredibly approachable and understandable, and yet you talk to her about the way she does it, she’s very clear on how it has to be statistically valid and how everything has to be in proportion and so on, So I think her work is great. I really like, on the more designer end, people like Bremer, who does visual cinema and does amazing work. She just did a project with us which was very complex bubble charts, but they’re all hand drawn, or they have a feeling of being hand drawn, and stuff like that but it took her a long time to do. But it was incredibly powerful, and visual and easy to understand.
Pennington: You’re listening to Stats and Stories and today we’re talking data and data journalism with Google data editor Simon Rogers. Simon I was reading something earlier you had from 2012, where you made an argument on data being the new punk, and I was wondering if you would talk a little bit about what you were trying to say and whether or not you still feel like data is the new punk.
Rogers: So every year now, about January, I get a whole bunch of data journalism students emails saying my professor said this piece about why’s data the new punk. So yeah I do still believe and I’ll say why, so the argument I made, and I was using a phrase I thought would resonate, and it came from Joe Cocker who famously said “anyone can do it” about his music. And I had seen this diagram that was constantly tweeted around, and it was from the punk fan scene from 1976 and it shows, here’s a chord, here’s another, here’s a third, now form a band, right? And there’s something about that, that ability to do this, which I thought was really powerful. So when I was first started as a working journalist, the idea that I could make a visualization that would be interactive- I could show – say there were 3,000 counties in the state, and I could make a map that would show the election results from all those counties, and I could do it, I wouldn’t have to ask anybody I just have to get the data upload it, make sure the codes are in there and it would work. That would be unimaginable, you know, ten years ago. And here we are with an incredible amount of free publicly available tools that anybody can use to make incredibly sophisticated visualizations. And you can see examples of this out there that’s inspiring to people who are- you know there’s a guy who did this project and it’s about weather and wind maps and he’d seen the hint from US wind maps, the global version I can’t remember his name now, of course, but it’s an incredibly beautiful visual, but this guy taught himself how to do it. So, for me, that in a way was one of the most exciting things about data journalism, how quickly it was expanding with people, some of whom were great, some of whom were producing that wasn’t great, but it kind of doesn’t matter because it’s like getting that stuff out there and getting people excited and enthusiastic about data is really important. And now, one of the things I do is work with the data journalism awards, and you’re increasingly seeing that entries from all over the world, like in places, like as one person in Afghanistan with you seeing really interesting data tools and using free tools that are out there, and just having the ability to do that. So, I guess I’m saying, in a really long-winded way, that I do think that this idea that anyone can do it is still really valid, and often is valid partly because in newsrooms data journalists are often on their own. There’s often one person working in an environment where people don’t understand what they do, they don’t know how to really support them properly, and they do it and they work together and they do it and they find the best way to tell that story using the incredible array of tools we have now. So yes, I do think it’s the new punk.
Campbell: So, when you’re teaching- and this goes to the math anxiety thing- and when you’re using this punk metaphor that anybody can do this, there’s some students who are going to be blocked, is there a technique you use to get them over that hump to show them anybody can do this?
Rogers: My course is a bit different from other people that teach data journalism, and I’m not saying it’s better or worse, to be honest it’s the only one I can do. I do not spend a lot of time teaching individual tools. I work with other professors, proper teachers who will spend the whole time teaching Sequel or something, and I think I’ve got like ten weeks with you, the thing that I can do is give you a sense of what’s possible and then help point you in the right direction. So one of the things we would do in every class though, is these kind of recurring exercises, one of which is look at the news stories that are in the news that day, go and find some data about them, work out what questions you want to ask, and then find out the answers to that questions, and do that every single class. So they’re always thinking how the data is applicable to something that’s going on in the world so it feels like it’s real, as opposed to this abstract concept, and I suppose secondly just to show how straight forward it is to find data, and how easy it is to visualize it or so stuff with it so I’m really conscious of it because that was me.
Bailer: That sounds like a great class. I have a quick follow up, so from a statistical thinking perspective not all data is equally good. There’s questions about representativeness and how is the data obtained, and to what population does it apply? So, there’s issue with a sample then, is it appropriate inferring anything? And another related question is a lot of times with visualizations it’s almost always just single numeric estimates without any sense of uncertainty.
Rogers: One of the examples I use is- there’s a journalist called Michael Blastland, I’m not sure he’s still there but he wrote a piece about the norovirus a few years ago. This is about norovirus reporting, and basically the reporting was all around how there could be three million people with winter vomiting sickness, which sounds like a lot of vomit. [laughter]
Pennington: It is, I think, in theory.
Rogers: Actually, this was entirely an estimate based on like 30 cases. And actually, it could just be 1,000 cases, it could be 500, that is all there in the data, it’s just hidden in the footnotes. And so, we spend a lot of time in my class looking at footnotes and looking at how meaningful something is. But trying to use examples like that which are not making people feel l like they’re in a crazily in-depth math class, because I want to be able to explain stuff properly and I’m not – I mean I recognize my limitations in these areas, I guess I should say.
Bailer: But that kind of diving into the footnotes to understand, I often wonder about who’s reading it other than your class?
Rogers: Well I think I feel like things have got a lot better.
Bailer: Ah, good good good.
Rogers: Because I have news on the whole time, because even though I’m now at Google I still treat with my team like we’re very much in a newsroom environment, so we always have news on. There are reporting about polling talking about the margin of error and talking about things like that that when something is one percentage point ahead of somebody else, it’s meaningless, right? Or they’re ties. And things like that are important to teach people, because they think there’s a certainty. You know, right now I work with data that’s incredibly uncertain. If you look at the way search data changes minute by minute, it will tell you 14 contradictory things in that time. And teaching people the uncertainty in the world I think is really important.
Campbell: So, one of the things that I was interested in was, you made a move from newspapers to social media sites, so you work for Twitter and Google, what prompted that decision?
Rogers: It was a couple of things. On one end I really wanted to live in San Francisco, honestly, and I thought it would be great for the kids to experience something else and those are kind of very personal reasons, and the opportunity came up. And the second thing was really around just the work. Because I really realized that I needed to work on how I dealt with bigger datasets, like how would you manage this stuff? Like there are billions of Tweets saying, “how can you tell a story with that?”, and I realized that would be a challenge for me, and so when that came up it seemed like a really good opportunity to do that. It seemed like that was really the move I was incredibly anxious about it honestly. At a personal level I was taking my kids away from everything they knew, and my wife and everybody had to readjust to this environment, which they all did way better than I did. [laughter] To be honest, there are similarities to being in tech company and being in a newsroom, that they are both dysfunctional families. Big personalities, and open offices and all those kinds of things. So actually, it was less different than I imagined it would be, but also, I realized I had to find my way; I had to almost start again a little bit. So, there was a bit of that. When the opportunity at Google came up, that felt like a better fit for me because it was really around working with news. I work in the News Lab, which is the team at Google that is really kind of a bridge- an editorial bridge for journalists and the company, which means we do editorial projects, and I get to think about what’s next, how things are going to change, how does technology make a difference in how journalists work? And I get to work with this dataset which is probably the biggest and most important journalistic dataset out there right now, and the one we understand the least.
Bailer: Very interesting. So is part of your work at the Google news initiative, what’s been the most interesting story that you’ve covered? Or what’s been the thing you discovered that surprised you the most?
Rogers: So, there are three things about Google data which- one of the things is that people associate Google with data, but nobody really knows what that data is. Three things that really hit me about the data- and I work with Google Transdata primarily, my team does, and Google Trans is a random sample of all searches. And you think how many searches there are a day; there are billions of searches out there, and what we do is try and find the signals in there. And the first people say oh it’s just a search, what’s a search mean? And I honestly don’t know the answer to that, it is a powerful social signal that we just starting to understand. We’re learning how to access this data really the last few years. So, the most obvious thing about the data is it’s big, there’s lots of it, and that hugeness kind of takes you beyond the echo chamber of social media into a world where it’s kind of ubiquitous. And that’s really powerful. The other thing about data is it’s incredibly honest. You’re never as honest as you are in a search engine, so it gives you an real view about what people honestly care about, because when you search for something you’re not presenting yourself, you’re just naturally inquiring about something you’re interested in. And the other thing is that it’s immediate. When something happens, there’s an instant reactional line because people want to know what’s going on there. And so, they take steps to it. So those three things. So, I have friends- so I wasn’t really expecting immediacy, to say I didn’t know it would be like that. Just the complete honesty in the way people search. And sometimes you see that reflected in the way people search like food, or life, or like searches for irrational fear of trees went up 75% last year. So stuff like that, which is just weird, or some things you kind of knew but you didn’t know until you see them in search, like searches for how to get to sleep are highest around the world in England on Christmas Eve night.
[Laughter]
Rogers: Or when people search for cocktail recipes which is the 31st of December, and the next day people all search for hangover cures. And this is how people search so those things are really amazing and interesting. You know we’ve done a lot of work around elections. I joined Google just before the 2016 election, so seeing how people search is really interesting, and seeing spikes in searches in places like Wisconsin and Pennsylvania for Donald Trump ahead of that election- in a way that was surprisingly high at the time but now makes complete sense.
Pennington: That’s interesting.
Rogers: Yeah, those things are really fascinating to me, and even though we’ve carved out into it and we’ve got a good sense of what you can get out of data now, I do feel like we’re still at the very early stages of fully understanding the significance of it.
Pennington: Well that’s all the time we have for todays episode of Stats and Stories. Simon thank you so much for being here.
Rogers: Thank you for having me.
Pennington: Stats and Stories is partnership between Miami University’s Department of Statistics and Media, Journalism and Film and the American Statistical Association you can follow us on Twitter or Apple podcasts or other places where you can find podcasts. If you’d like to share your thoughts on the program send your emails to statsandstories@miamioh.edu or check us out at stantandstories.net and be sure to listen for future editions of Stats and Stories where we discuss the statistics behind the stories, and the stories behind the statistics.