The history and business value of data visualisation | Interview with the Head of Insight Analytics at the FT

PODCAST

Feb 2021

Arrow Icon

What's the history of data visualisation, and how can companies use it to effectively display their insights? We interview Lucy Alexander, Head of Insight Analytics at the Financial Times, to find out more.


You can stream or download below:

Lucy_A_Podcastimage.original.png

Lucy Alexander

Jenny Stark

Jenny Stark

Lucy A Podcast
Loading...
0:00 0:00

Podcast Transcript

Jenny Stark: Hello, and a very warm welcome back to the FT Strategies podcast, bringing you the latest strategies for digital growth in the business world. I'm your host, Jenny Stark, and in this episode we're going to be talking about data visualisation: how companies can visually tell stories using data analytics. Here joining me today is Lucy Alexander, Head of Insight Analytics at the Financial Times. Lucy, thank you so much for joining us.

Lucy Alexander: Thank you very much for having me.

Jenny Stark: Let's jump straight into it. What is data visualisation?

Lucy Alexander: Data visualisation is a means of displaying, in visual form, data that you are interrogating (and have interrogated) so that people can understand it clearly in that way. A data visualisation could be as simple as a bar chart, it's a very effective way of displaying data, or it could be much more complicated as we've seen with some of the data visualisations that the FT has published along the coronavirus lines, where we're talking about rates of infection in different countries and how they're spreading over time - these are really sophisticated pieces of information that we're trying to convey. Doing them through a visual means allows people to absorb them and really look at them and take them in and compare in a way which probably wouldn't be so easy to do in any other way.

Jenny Stark: What are some famous examples of data visualisation?

Lucy Alexander: Data visualisation is something which has been really thought about and considered for centuries, actually. When you think about the cholera outbreak in London...it's been written up in a book called Ghost Map which effectively outlines how they first...I mean, this is a really grim subject sorry but I feel like I'm so overwhelmed by covid thinking that stories about other outbreaks are always top of mind! It's how science first understood that cholera was a waterborne disease, and they did that by essentially a couple of really unbelievably scientific men annotated on a map where all the people who had fallen ill were, and how they had behaved in their final days. And they realised they were all grouped around a particular well in the City of London - actually I think in Soho. And they were then able to say, well it looks like this well is the cause of this illness outbreak. And they were able to trace...there were individual cases that were really far away but they were able to track these people down and say, what were you doing in the days before you got ill? And it turned out that they'd all drunk from this well. It was just an enormous step for science and for medicine. There's a beautiful visualisation which accompanies it which is essentially just a map of that particular part of London with the individual people who had contracted cholera as little dots around it, and the picture it paints is so incredibly obvious. You know, they're literally just circling this well.

Jenny Stark: And what about modern examples of data visualisation?

Lucy Alexander: Things as obvious as the London Underground map. Those distances are not right...it has been reduced to very simple component parts. Each line is represented by a colour and it's made to be as visually understandable as possible. I personally love to geek out on these things. There is a real art around the kinds of data visualisation you see around election time - how people display maps to make them immediately clear and when you're understanding about population size voting versus the electoral college system for example, there are ways of displaying that to make it clear what is the nuances behind the story. The coronavirus briefing that many people watch whenever it's on always has charts, you know, it's always full of data visualisations. This is the clearest time I've ever seen when charts have been absolutely in the mainstream and we're all looking at the same data and trying to absorb it. So I think this is probably the most pertinent moment to bring up the power of data visualisation in trying to explain a concept as complicated as an airborne virus to a nation who are fairly tired.

Jenny Stark: Does the FT have a method for its data visualisations?

Lucy Alexander: How we try and approach it is, we come with hypotheses in advance about what the story might be, and we try and either refute that or support it with the data that we have in place. And we tend to come with a very open mind, so we don't mind which way it goes and we try and look at it as objectively as we can. But the data is effectively your kind of, your individual dots on a painting and it's assembling them in the correct order that the picture reveals itself. There's another element to this which is essentially how you communicate effectively, how you tell a story effectively, and that's a completely different skill, and so requiring analysts and professionals to be good at both is really kind of quite exacting, but it is so effective once you're able to tell in really interesting narrative form something which relies heavily on data because you've got that objective core to your story.

Jenny Stark: The FT has a pretty good reputation for its data visualisations.

Lucy Alexander: It's well known that the FT is absolutely outstanding in this area and John Burn-Murdoch is often touted as being a real master in this; he's senior data visualisation journalist so he sits within the editorial floor of the company and, you know, the work that he has done on the coronavirus tracker is frequently touted as being absolutely outstanding and it is. Because's it's clear. Because he's really thought about how you display the data in a means that's easily understood. You know the charts are clear to read, the colours are all differentiated, he's annotated them very clearly and that's an important point, is to always remember, you know, does this chart contain enough text? Because there's a real marriage between the actual text you add on top of a chart and the actual visualisation of a chart, the two go together. And making sure the words on your chart are really unlocking any mystery that might be there, that you're pointing to significant things and de-mystifying what they are - all of those component parts are really important.

Jenny Stark: And is there a ruleset for how you approach and deal with different types of data?

Lucy Alexander: The consideration is so important in terms of what is the right visualisation for not only the story you're telling but also the specific piece of data you're trying to describe. There's something called the visual vocabulary. We have our own at the FT and there are many you can see online, you can Google visual vocabulary and you can get examples of this, where there is an appropriate data visualisation for each different data point that you're trying to describe. So for example, if you were looking at comparing magnitudes there are several different ways you can do that really effectively which is both clear and also accurate, so you're able to tell that story in a way which is both objectively getting the information across in a really effective way but also looks really clear and you're able to digest it easily in your brain. It also helps you to steer away from some common pitfalls. People have traditionally really overused pie charts for example, and pie charts have their place but it's really important to consider, is that the clearest means of doing it? If you're trying to compare the size of many different elements then probably a pie chart isn't the right one for you and using a visual vocabulary that's already been laid out quite clearly by industry experts such as our own Alan Smith is a really easy way you can try and hack in and not make any mistakes when you're thinking about how to present the data that you have.

Jenny Stark: Is there a step by step method at the FT for data visualisation?

Lucy Alexander: There are so many different elements to coming up with a finished product where you've got clarity of understanding and you've got clear recommendations of what you should do on the back of that piece of insight, and those steps along the way are really important from the point of view of making sure, firstly that you don't introduce any errors, secondly that you've really thought through the implications of the question and you've fully understood the brief. Making sure that when you're actually getting the data and you're wrangling the data that you're doing so...you're using the right methodology for example, you're not introducing any bias. So we have guides for our younger analysts to make sure they're following those processes as best they can and we have QA steps in place where we try and make sure that if any errors do creep in we can capture them. And then finally there is another layer of QA-ing which is what are you trying to convey with this particular story or chart? What is this analysis trying to actually tell us and have you actually managed to do that - like, is that clear the message that's coming across? And for that you tend to have to have a slightly more broad view of the company, a little more experience in these things just to make sure you're not just relaying the data but actually demystifying the data, explaining why it's important and why it's relevant.

Jenny Stark: So, do you come with an agenda for what story you want to tell and find the data for that, or does the data tell the story?

Lucy Alexander: Essentially the data is the story and it's about how you try and get into it and how you unlock it and how you wrangle it, is actually you just finding the story in there. So, used effectively the data will tell the story itself. The data is really just the sum of all the actions and you have to try and look at those actions and try and interrogate what the purpose behind them was, what was the intention, and from that extrapolate the implications for the business.

Jenny Stark: What's the business case for using data visualisation?

Lucy Alexander: Data visualisation is a weapon in your arsenal. There is an enormous argument for why understanding data and being comfortable with it and really embedding it into your thought process is good for your bottom line. Would you ever feel more confident to really go with a big decision if you had underlying data to support it or if it was entirely led by instinct? It's obviously much more comfortable to use sound layers of expertise in order to base your opinions so you not only mitigate your risk but you're also opting for an option which has some proven quality. The more you get comfortable using data in a sophisticated and thoughtful way the more likely you are to make good decisions I think. That's my opinion of course, I'm biased! But I think there's a lot of evidence to prove that's the case.

Jenny Stark: What kind of tools do you recommend businesses use when handling data?

Lucy Alexander: You can get a very long way along the journey just by using the basic tools you'd expect most companies to have, things like excel or Google sheets. They're not incredibly sophisticated but they're also capable of delivering something which is pretty good. There are other - I mean, not wishing to plug people specifically but there are companies like Tableau who are well known in this space who make doing really beautiful data visualisations quite easy for people. If you are at the very sophisticated end of this and you're using coding languages, so if you're using R for example as a routine way of doing your analysis, then you can do really interesting and thoughtful data visualisations using those as well. So, you can start absolutely at the basics with the kind of packages that would typically be on most people's laptops anyway, and go all the way up to something sort of really quite exceptional. But like I say, I think it's pretty widely available to anyone and software like Tableau which sits somewhere in the middle is a really effective way of getting quite far along that journey.

Jenny Stark: At what point during a project do you recommend companies start using data visualisation?

Lucy Alexander: I think that the data storytelling should start at the beginning of a project, personally, but it's probably easier to come at a data story in the way that we kind of understand it, I think most typically when there's already quite a lot of established information, because of course you then sort of have a sneak preview of what the narrative is essentially going to be. At the beginning of a project it's much harder to visualise what that's going to be. I always think about starting...you have already got some of the information there, like you already know why you're starting the project. So let's assume that we're going to be collecting data on a product that's being released. You know why you're doing that. So you've already isolated that there's probably a gap either in your offering or in the market that you're trying to fill with the product, and you've got a hunch that the solution that you've found is going to be really effective because you've already shown you're going to speak to customers and you've done some thinking about it internally and you're pretty robust in your thinking. So that's already the introduction to your data story. And then your story will evolve as your project evolves. So you may produce the first prototype of your product and release it and get some information and gather some data about how it's doing. You've then got the first data points you can use to really embellish what you already thought was going to happen. Like, did it support your hypothesis? Are people responding to it in the way you thought they would? Or has the initial release maybe not gone so well and does it need some iterating on? So there is an interesting narrative already developing here, it's already quite complicated, and I like to think when we're doing projects along those lines so that the narrative is this kind of evolving thing.

If you're further on than that and you're coming to something which does not have a data story for example written about it yet, but it could be something like your North Star metric, and you're trying to make sure you communicate that to the company and get people to understand what that is, the important thing is to remember what is impactful about narratives in general. You know, storytelling is as old as humanity, it's the way we communicate messages to make them memorable and you should try and incorporate all of those techniques you know. Is there interest in it - where is the drama? But with data it's always important to remember you're as objective as possible, you know, weave these facts that you actually know through the data and into this narrative. It's really important to be credible and it's important that you're communicating really clearly because anything with data - there's always the danger that people are going to be lost along the way, because people's level of comfort with data is very varied. So I think it's making sure you're communicating it as an interesting story, which it is! From a human point of view mostly, making sure you're keeping it always on track and keeping it as objective as possible.

Jenny Stark: And when should companies use data visualisation?

Lucy Alexander: They should be using this kind of technique all the time. I mean, the idea for an effective company I think would be to use data as a means of directing decision making on a constant basis. It doesn't mean you have to be writing a lengthy data story. We actually do produce things called data stories at the FT, we have an outstanding human called Jono who writes absolutely beautiful data stories about ongoing projects and metrics of importance and things like that, in order to unlock those for audiences who perhaps wouldn't be aware of the detail, but we should be weaving our data into our wider narratives all the time. At the FT, the journalists are experts at doing that. So we have people who are absolutely state of the art at telling stories and weaving data into them and making data central to them. So covid is a perfect example. All the time we're consuming understanding of the situation of the covid pandemic through the means of data. I mean, we are now sitting and waiting for the Prime Minister to tell us everyday. Data points are then getting plugged all over the world across different news stories and the FT is doing that particularly beautifully. But in an internal sense there are stories evolving constantly. So every product that's being produced essentially has its own narrative and you should be weaving data into the narrative in each of those products as they go along, it's an ongoing evolving process.

Jenny Stark: Any tricks when it comes to presenting data for stakeholders?

Lucy Alexander: I think when it comes to communicating data effectively to the business you have to think of the context of that particular question. Is the stakeholder ultimately really just interested in understanding which decision they should make? So in that case you may not need any of those methods, you simply need to put that in a paragraph, we believe this is the best recommendation for you for the following reasons. If you're looking to illustrate a point in a way which is going to resonate and hit home and have people really mull it over then we would always try and include a visualisation. You know, it's something people can really get to grips with and really understand the concepts involved and it's an effective way of conveying the message in a way that a table really never can do. You're visualising it for people and many people are inherently driven by their visual learning rather than anything else and anyone who is uncomfortable with maths, you know, not everyone pursues maths as their degree or carries on doing that indefinitely and many people stop at the earliest possible moment and are very happy to do so. So you don't want to put those people off from really getting to the heart of the matter by including too many tables in your thinking if they're not necessary.

Jenny Stark: What's your predictions for the future? Can you see a world in which data visualisation is embedded in our reality?

Lucy Alexander: There's been a really beautiful resurgence of the public in general wanting to consume data visualisation in the last decade to two decades, and books have been coming out which have been extremely popular on the subject and really just emphasising the power of the beauty of it. So I think that there's no doubt as we move further and further into this digital world, and particularly as we're all remote now, it's such a big part of our lives, we're probably consuming more of this sort of thing than we ever have been. In terms of, will we be entering a separate reality where data visualisations are embedded within our virtual reality world, I'm sure this is full established! I mean, this is not my area of expertise but I'm sure within the world of video games and things like that that these things are happening already all the time and they will only continue to happen as we proceed down that path.

Jenny Stark: Any final book recommendations?

Lucy Alexander: Yes...so, the Data Is Beautiful is absolutely outstanding...there are so many, it's really worth having a quick Google and seeing which ones are visually appealing because a lot of them are explaining what the art behind the art and science combination is, you know, it's really meant to be stunning and I think it's subjective as to what you find personally appealing so I think it's worth having a look and seeing what's out there. But there are absolutely loads of great texts on the subject now and everything from really huge enormous giant books to things which are much more easily digestible. But there's so much out there and there are real experts in the field that you will quickly tap into who have their own visual vocabulary and their own really distinct way of doing things and it's beautiful.

Jenny Stark: Lucy Alexander, it's been thrilling talking to you, thank you so much.

Lucy Alexander: You're welcome, thank you so much for having me, it's been a joy for me as well.

Jenny Stark: That's it for this episode. To listen to more FT Strategies podcasts or to find out more about FT Strategies, please visit www.ftstrategies.com

Report on the Google News Initiative & FT Strategies Subscriptions Lab 2020

Google: Subs Lab Europe

Get started

We've helped companies around the world future-proof their businesses - and we can do the same for you.


There's a lot at stake when you're creating a digital revenue stream - so naturally, you want to arm yourself with the best knowledge and expertise to make the correct decisions. Our team bring over 100 years of consulting rigour and world-class FT digital experience, to help you thrive in the new digital economy.

Get started