Ed and Steve discuss Kenneth Field's recent blog post Mapping coronavirus, responsibly.

The Geomob podcast is hosted by Ed Freyfogle, co-founder of OpenCage, and Steven Feldman, of KnowWhere Consulting.

Every week we discuss themes from the geo industry, interview Geomob speakers, and provide regular updates about our own projects.

Autogenerated Transcript:

Ed 00:01 Welcome to the geomob podcast where we discussed you innovation in any and all forms good for fun or problems. Welcome back everyone to another episode of the geomob podcast. Today I again have cohost Steven Feldman with me and we're doing kind of a special episode today talking about mapping the Corona virus, but firstly, let's catch up on what we've been up to. Steven, how's life? What are you up to?
Steven 00:28 Well, apart from the fact that everywhere is gradually closing down because of fear of coronavirus, I'm pretty good at healthwise. Fine. London is still plodding on. Cubes are absolutely rammed packed, so beatings are going on. Nothing seems to have come to a stop yet, but we'll maybe talk about that in a little bit. I did have some fun last week wrestling with SSL certificate and I've just wanted to give a shout out for anybody who needs to use SSL to the let's encrypt service, which is a free service generating SSL certificates. Use that with something called zero SSL and I'll put the links in the show notes with you. It's a really great service. I generated six SSL certificates for various websites in just under 40 minutes, no charge at all. So it's a great service for people running their own websites who want that. The other thing that I've been doing recently is, you know I meant a CEO was part of my consulting business and I had one where I had to break, which was really great for the people that I'm working with. You know one of them was a really simple thing about tracking some key stats on a weekly basis and just taking 10 15 minutes at the end of the week to pull up your key stat, you know, some order intake figures, maybe some billing figures and just look back on the week and digest how you've done in the week. And the other one was
Steven 02:04 introducing a new CEO to -inaudible- the concept of the value proposition canvas, which is a great tool if you're trying to work out how your product fits in different markets or understand what your customers,
Steven 02:20 and I'd recommend that to anybody who is in product management, product design, and in fact CEOs running small businesses who just want to have a single service that they want to align better with their customers. So
Steven 02:37 I've had a good week.
Ed 02:38 Well congrats actually that the topic of mentoring, I know you did that with quite a few startups but also geo startups. So maybe that's something we should go deeper on in a, in its own episode in the future. Probably be interesting. But what do you, well it has not been the most productive way cause we've had my, my sister in law visiting, so doing some touristy stuff. This is, this is one of the big risks of life in Barcelona is that a lot of people like to come visit, which is great, but you end up playing host and doing all the touristy stuff and it's great. But as a result, at some point you look up and you're like, what have I done for the last six months? You know, it can be, you know, just always hang out at the beach. But it hasn't been a lot of fun.
Ed 03:19 I'll not least because we've had great weather here. The spring has sprung. So it's been nice. Well, as you can say, it's a bit of paradox in that the, the weather is so nice. Spring is here, spring is in the air. And meanwhile of course we have all the bad news about the virus and events being canceled and it actually, I'm kind of pondering, wait, what is the implication of this for geomob? Obviously do your moms are relatively small event and I don't think many people are traveling across international borders to come to geo, but, but still, you know, is it, should we be canceling geo Mav? Should we not? Is that an overreaction? Not sure.
Steven 03:55 I'm not sure either. I mean, I think, uh, certainly at the moment London seems to be carrying on as normal. But you know, every day we're becoming more and more aware of how serious this this virus is. And you can see things, you can expect, things will close down. You know, we're a couple of weeks out from -inaudible-. And the good thing for us is that we could make a decision 24 or 48 hours in advance.
Ed 04:23 Yeah, well you can, I actually do travel to London for GMR, but I guess for the benefit of our listeners where we're recording today on the 5th of March. Yeah. And the next London geomobs on the 18th and then we have one the week after that on the 26th and Munich, which I was also planning on attending. And actually I would, I would be heartbroken if I couldn't make it to a, because I'm speaking, but B, you know, for the last year, geomob, Munich has been running quite successfully. And for whatever reason that's always been a conflict in my schedule and I haven't been able to attend. And this was the first one I was finally going to go to. So I'm really hopeful. I'm still able to so,
Steven 04:59 well let's just keep our fingers crossed. It's difficult to predict day to day what's going to be happening, but perhaps that takes us onto the topic that we thought we chat about this morning, which is the way people are mapping the coronavirus. Shall I start by just letting,
Ed 05:17 yeah, why don't you briefly mentioned Ken's article which kind of triggered this whole discussion. Yeah
Steven 05:22 and I think the first thing we should do is say that actually there's you and me talking about this and it would be a heck of a lot better if one of us had thought to schedule a call with Ken and we'd got Ken on to talk about this or maybe in a few weeks time when things are settled down, we can get Ken on and he can talk about mapping viruses and mapping elections and things like that. Cause he's got lots of interesting stuff to say. But with that in mind, I'll paraphrase what 10 let's say for those of our listeners who don't, Ken field is a cartographer. He works as re, he's written what is considered by many to be the topic leading definitive book on cartography. Yeah. He's a world renowned cartographer and he cares passionately about making good maps and what's happened is that with the outbreak of Corona virus, the media have been publishing some of the most horrible maps that you can imagine on the subject and every mistake that a first year cartography student would have learned how to avoid these guys are making, and Ken started off, he wrote a series, a fairly long thread on Twitter about the subject, and then he went on to write a blog post, which the link will be in the show notes.
Steven 06:54 You search for mapping Corona virus responsibly. You'll probably find it very quickly. Basically what he's saying is, yeah, if you're going to map this, and I keep, I want to keep saying if you're going to map this because I think we should have a conversation about whether you should map Corona bars, but if you are, well, let's take that at the end if that's okay. Yeah. If you're going to map it, 10 suggests there are some basic rules that you should follow. The first is you should not use the Mercator web map projection to map coronavirus. You should use one of the equal area projections that it gives you a more balanced view of the size of different countries because one of the problems when you start making a clearer path map where you shade the map to different colors depending on the intensity of the variable, is that visually the big countries always dominate the small countries.
Steven 08:06 And the second rule that is, which it's so obvious that it's astounding that people so do this automatically is you should normalize the values that you're mapping. And in this case we're talking about incidents of Corona virus or deaths from Corona virus, and if you take the country is big as China, which has got a population of nearly 2 billion I think, and you've got 3000 deaths, right? If 3000 is a massively bigger number than any other number on the map and it's spread across 2 million people on a massive geographic area, you get this big typically red blob where China is, which is totally misleading. So what you should be doing is mapping incidents or dare incidents per hundred thousand or per million of the population. And you should be mapping deaths either as proportion of the population or as a proportion of the incidents, you know. But either way what you need to do is normalize
Steven 09:26 the data so that the -inaudible-
Steven 09:27 distortions of large populations and small populations are eliminated from the evidence from the map that you're presenting.
Ed 09:36 I think a lot of this advice though is not necessarily specific to coronaviruses. I mean it's not specific to any type of not being actually by coincidence. My wife is actually an epidemiologist, although she doesn't work on infectious disease, but you know, she, she will always complain about how, you know, various complex scientific topics are misunderstood and misrepresented. And yeah, as you say, like, you know, the numbers are not taken in context in terms of to the, to the proportion of the population and things like that. I mean, this comes up time and time again.
Steven 10:08 Did it does. Indeed it does. You know, and I mean, Ken's not saying anything that he hasn't said.
Steven 10:14 I've given talks on this subject and said the same things and he's written a book about it. It's just that coronavirus has generated this enormous amount of media and work with it, a colossal number of maps.
Steven 10:29 And actually that's a good point to actually talk about. Why do we think there are so many maps of coronavirus at the moment?
Ed 10:38 Well, I think this gets to the point of whether our map even makes sense or not, and we're talking about this a bit before, but the purpose of a map is you look at it and you, you gain understanding about, and you gain a certainty of, as I'm, you know, where, where is the stuff? Who owns it? How am I going to get from a to B or whatever. And you know, what we see with coronavirus is the way people move nowadays with airplanes and trains and stuff. You know, it's not as if it purely just spreads geographically. Like you know, one day it's all in China and then all of a sudden you know there's a user explosion in Italy, which is obviously on the other side of the world. So you know, I think people are using maps, news organizations are using mouse because people want to have certainty about this uncertain situation. But I'm not sure it's the right tool in that regard in that it doesn't adequately represent the way people are moving the flows of, of, of, of humans and thus and also the virus.
Steven 11:33 I think that's very true. And I think if we look back to the beginning of coronavirus we had, and it will be interesting if somebody collects a timeline of maps so that we can look at the maps from the beginning of this, this incident and the
Steven 11:49 maps that have appeared later on. Because
Steven 11:51 the beginning, I, if I correctly, we were seeing maps where basically there were big red dots or whatever in the center of China in who way province. Right. And if you were sitting in London is no Barcelona like we were. You're looking at that and you're thinking that's a long way away and as long as not too many people get on a plane from China, we should be okay. And I think what we've seen is obviously as you said, people, you know people have got on planes, people who've traveled, the viruses spread. And I agree with you that I don't think the map is actually telling us very much. Yeah, there's probably, I think there's 70 countries in the world now day, but I've got coronavirus seeing those countries lit up in some way with small,
Steven 12:40 some big dots doesn't really do anything.
Ed 12:43 Well, I mean this is, this is going to be like,
Ed 12:47 yeah, it's going to be like the cartoon a, I forget the exact phrasing, but basically the point, the point of the cartoon is just that when something that's proportional across the population event, eventually you just end up with a map of population density. Correct. Yeah, and it does, it's not going to, you know, especially something like the virus which seems to be quite infectious and spreading quite rapidly. You know, basically it's going to be like, Oh more people in big cities have the virus than out in the country. Yeah. Obviously what are, what are we learning from that? Yeah,
Steven 13:19 and I'd say I think you know, there is a real case for saying that the maps are not helping at all and if they, if they give a full sense of certainty in any regard, whether it's a sub D that is negative or positive, I'm not sure they're helping. I think, and I actually recall that at the end of his article can ask the question whether we should be mapping this at all and suggested that a bar chart might be a better way of representing the data.
Ed 13:51 Hmm. Yeah.
Steven 13:52 Yeah. Because
Ed 13:53 yeah, I mean, I, I do think this raises some interesting issues in general about how, how can news organizations, um, which obviously in recent years have been under massive kind of financial strain. As the old models, old business models fall away. And so, and also new methods of spreading of information via social media. You know, social media posts can literally go viral as the term is, you know, there's huge potential for inexperienced people to but well intended people to, to make representations of data that then spread wildly, but, but can they wrong information or, or, uh, you know, lead to panic. And, and the question is, what should organizations do about this, um, or news services? Well, I mean, I don't think training everyone to be an expert cartographers necessarily going to be the solution or the set of realistics realistic.
Steven 14:49 Yeah. I don't think it's unreasonable. Expect a large
Steven 14:52 media organization to employ some experienced cartographers.
Steven 14:58 I agree. I think that is true by the BBC. For example, we had a talk in London that's one of the longest
Steven 15:06 Jude events from the BBC cartography team and it was a really good talk and it was really interesting. And they work under enormous pressure, you know, because they're working to news room deadlines to get stuff produce and they do some amazing work. But they've also produced some of the most diabolic maps of this Corona.
Steven 15:29 Um, so yeah, and they do employ people who are relatively skilled. They may be
Steven 15:35 more GIS people than, than they are cartographers. But they do employ relatively skilled people
Ed 15:42 at open cage. One of our customers is a service based in Germany, but available in multiple languages called data wrapper. And specifically it's a tool for news organizations to build infographics, be one subset of that being maps, but also charts and things like that. And I think maybe that's the solution is that you have kind of a best of class tools that help guide you to, you know, you kind of, I, you know, I don't know how you do this from a usability perspective, but you come into the tool and say, okay, I want to represent this thing and here's the data I have and maybe a guide you to like, you know, in, in this situation, this type of representation would be appropriate. This one would not. And here these things to look out for and you know, should you be, you know, make sure you're, you know, maybe raw numbers are not appropriate and said you should base it on as a percentage of the population. Things like that.
Steven 16:32 Oh nice.
Ed 16:33 And helps push you towards the best practices.
Steven 16:35 Well I haven't used them for, so I'm speaking
Steven 16:40 at a slight distance, but I do believe actually that some of us are, his tools do go some way down the line to doing that, you know, suggesting that the right color palette.
Steven 16:51 Yeah. And stuff. Because one of the biggest problems is making a map using the default settings. And I think another problem that we've got, which may surprise you, is we've made it so bloody easy to make a map.
Ed 17:07 It's true, it's true. Well this issue of the default settings, I mean this is something that comes up with every tool, right? Who is the audience? Is it the advanced user? Like you know, Ezra is targeted more at advanced users with kind of a GIS background or is it, you know, data wrapper, that's what's good about their service. Targeted specifically at journalists who perhaps don't have, you know, full geographic GIS training or anything like that. And you know, probably as you say, they're working under a very tight deadline and they need to just get something out quickly because you know, if you're not the first one on social media than someone else is. And
Steven 17:43 I think also if your talking about the
Steven 17:47 media, Oh, so the website, the guardian website, whatever. Typically one of your objectives is to get people to spend as long as possible on your site. Go in. If you're the guardian for example, you want them to click on your advertising, you know, cause that's part of their business model and a map with lots of detail on it causes you to spend more time on the site, doesn't it?
Ed 18:09 Yeah. Although I, I'm going to hope, you know, maybe, maybe I'm too good natured, but I as a reader of the guardian, I don't, I don't think they're trying to make, intentionally make the infographic confusing as a way to get me to spend more time on the site.
Steven 18:21 Okay. All right. I'll, I'll back away from that one. But I think you're probably right. Forget I said that. Let's do it.
Ed 18:28 Well, I did this whole thing. Yeah. They'll go, Hey, good, good.
Steven 18:31 Just set out some guidelines for how you could make a reasonably responsible map of the virus. But in parallel
Steven 18:42 last week I saw a cartogram on Twitter. Do you know what a coach Graham is?
Ed 18:50 Yes, I do. But maybe you should, should explain it for the -inaudible-.
Steven 18:53 I'm going to give you just a quote on what a cartogram is.
Steven 18:58 Cardiograms offer a way of accounting for differences in population distribution by modifying the geography. In other words, they stretch the geography too, adjust for population differences. And so big countries, big populations, regardless of their -inaudible- geographic size get expanded. So England for example, which is a pretty dense country, gets expanded. South America would get shrunk and things like that, but you end up with
Steven 19:32 almost, I don't know whether they're like a doc,
Steven 19:35 if Salvador Dali withdrawing maps, you think they might look a bit like this?
Ed 19:41 Yeah, they're kind of twisted and pulled and stretched and other places are shrunk
Steven 19:46 and I looked at this thing and I thought, what on earth is that showing anybody?
Steven 19:57 Yeah. Even somebody like you or me who understands what a cartogram is doing is going to struggle to get any understanding from this thing. The only thing I can tell you is China becomes a big ball on the right hand side of the map. And I asked this question on Twitter and got back loads of feedback. Yeah. I said, what does a cartogram add in terms of insight compared with a normalized thematic map or even a bar chart. And strangely enough, Ken came back at me and said, why does it have to add something? It's just a different way of normalizing
Steven 20:37 the shape is normalized rather than the data I view is that cardiograms don't work for the general public. You know? And going back to what we were talking about a few minutes ago,
Steven 20:48 you were saying we can't expect everybody to be a cartographic expert. You have to be more than a cartographic expert to globs at a cartogram and extract much information from it. And one of the reasons I think that there's such a problem is that they distort the shapes to such a great degree. And my opinion, and it's not just my opinion, I found a bit of science to back it up, is that we have, we have a sort of visual memory of the shapes of the countries of the world. And most of us know roughly what the world map looks like and where countries are on that world.
Steven 21:33 And we can adjust that. You know, when we see the K to protection or uh, pizzas production or some other projection, you know, which slightly changes the say we can, we can cope with that. But when you distort the shapes to the extent that
Steven 21:49 the cartogram does where it's adjusting the shapes of the countries for the population, I think you come up with something that is unreadable by most people.
Ed 21:59 Wow. A lot to unpack there. See this, first of all, I'm shocked to hear that you think there might be information of no particular value on flowing about on a social media channel. So, but second of all, no, I think this gets to the, the crux of what it means to work in geo now is that we have so many different ways of doing things and tools and services. It becomes much more about always knowing what is the right tool for the job, right? And when to employ tool, Hey and when to employ a tool B and when to do nothing. And the problem is of course, if you're, you know, if all you have is the hammer, everything looks like a nail. If you say, well, you know, I really want to make a cartogram fine, you know, you can, uh, you can, you can do it.
Ed 22:46 Whether it adds any value is a different question. And I guess when it, when it's a topic, when I like the current virus of course, then it gets into is there a, do you have a burden of responsibility to not spread misinformation and spread panic and things like that. So, but really I think, I see this is the core of what the industry is now. It's, it's about knowing what the tools are and what the pros and cons of each tool is and when and how to apply them. And it's much less about, you know, the hardcore engineering or you know, the geo is just data cleaning and then using that data in the right tool to, to produce an outcome. At least that's how I see it.
Steven 23:23 I think there's a,
Steven 23:25 when we convey information, whether it be in a chart, in a, a map in some other kind of infographic. What we're trying to do is distill complex information, large volumes of data down into some simple graphical representations, which is allows us to explain something to the audience. Would you agree with that? Roughly?
Ed 23:56 Yes. I would say so. Right.
Steven 23:58 What we want to avoid is conveying the wrong information, you know, so using an anomolized map for example will convey
Ed 24:08 certain piece. You'll impression you can, you can easily mislead. Yeah.
Steven 24:14 It's not just about tools. It's also about having an understanding about the data that your working with and having -inaudible- an understanding about how you can convey information from that data effectively. And I think when your end goal just becomes -inaudible-
Ed 24:36 -inaudible-
Steven 24:38 or shading up a map, putting dots on a map or shading a map and if all you're trying to do is achieve is to get your data onto the map, then you are going fail pretty frequently in actually conveying information. And you probably need to understand the data that you're mapping before you start mapping it. And I think that's a big problem that people don't understand what they're doing. You know, if you asked an epidemiologist, if you, if your wife was mapping this, this data, I'm sure she'd have a different approach to it. To a journalist at the BBC.
Ed 25:13 Yeah. I'm not sure my wife would be the best a mapper, the long running joke in our family to how badly she reads them. But um, Oh, let's not go into the, uh, but I, but I do take your point. Yes, I do take your point. Oh, this is also what makes it hard. It's like exactly. You know, you need someone who has the expertise to understand the data. Then you need someone who has the expertise to how to represent the data. Then probably if it's to really get your message out, you need someone who needs to understand how to market it, uh, you know, or, or, or distributed. So it gets harder and harder and all those different pieces need to come together to create a successful,
Steven 25:58 where you've got to do that maybe several times a day. You've got to be updating this stuff. That's very challenging.
Ed 26:05 I agree. I agree. I mean, so I guess what's our takeaway from all this beyond the, everyone should obviously read Ken's article cause I think it's quite useful and thought provoking.
Steven 26:14 Yeah. It's not like it's a, it's a five, 10 minute read and uh, if you're making maps of the coronavirus or anything else, you know, frankly read this article because it's a really good basic primer in making semantic map. My takeaway is that we probably shouldn't be mapping this data. We should be using charts and tables and things that allow us to get a little bit more understanding of the data than the maps that we're seeing are providing. My second takeaway is I've never liked cardiograms and the flurry of them that I've seen recently hasn't changed that. What about you?
Ed 26:54 Well, I, so first of all, I agree, probably a map is not that helpful in, in an uncertain situation. Like the spread of the virus. Seeking to seeking certainty via via app is probably not the best thing to do. Instead, I encourage everyone to follow all basic safety and health precautions. And, um, and I guess we'll leave it at that. I, I do want to end things on a slightly positive note. I did see one funny thing going around on Twitter and that was are you familiar with this site where they, um, they have maps that leave out New Zealand? Have you seen that? Okay, so, so, um, you know, very frequently people will create a world map, you know, kind of centered on Europe and Africa. And as a result New Zealand, which would be, you know, far off on the right, it gets cut off. And so this guy collects maps that don't have New Zealand. And so going, you're on a Twitter, there was a map, you know, showing the spread of the current virus and you know, the bright red dots and everything. And, and someone commented, this is fantastic. We can't get sick if we're not on the map because the, you have left off New Zealand. So,
Steven 28:01 well that's a good, good note to finish on, Isaac.
Ed 28:04 Exactly. So to all of our listeners, you know, do your best to stay healthy and we'll see how things go with regards to geomob in a couple of weeks. I'm hopeful we can, we can still go forward and uh, yeah, please map responsibly.
Steven 28:17 Okay. Take care of it.
Ed 28:22 Thanks everyone for joining us today and listening to the geomob podcast. Hopefully you've enjoyed the discussion. Please don't hesitate if you have any feedback for us or any suggestions for topics that we should cover in the future. You can get the show notes over on the website, which [email protected] while you're there, if you're not yet on the mailing list, please do get on the mailing list where we once a month send out an email announcing future events, summarizing past events and just generally sharing, uh, events that you may find of interest. You can also of course follow us on Twitter where our handle is geomob. You can follow Steven at Steven Veltman. You can follow me fry Fogel, you can check out -inaudible- at -inaudible- dot work and of course, if you need any geocoding, please check out my service, which is open cage data.com. We look forward to you joining us again at a future episode or, and of course seeing you at a future geomob event. Hope to see you there soon. Bye.