FEATURE7 January 2013

Telling tales

Trends

There are many good reasons for researchers to embrace data visualisation. One of them is to make people pay attention to the story you want to tell.

Good design can be many different things to many different people. There are rules of course, do’s and don’ts to follow – the kind they teach in college. But ultimately, the question of whether a piece of work is good or not is down to the individual. It’s almost entirely subjective.

That’s true of data visualisation too. Whether a data visualisation does what it’s meant to do really depends on who’s looking at it and what they take from it. Most would agree on the definition put forward by Vitaly Friedman, founding editor of Smashing Magazine, who in a 2008 article stated that: “The main goal of data visualisation is its ability to visualise data, communicating information clearly and effectively.”

But, as Friedman noted, clearly doesn’t have to mean simply, nor does effectively mean immediate.

“To convey ideas effectively, both aesthetic form and functionality need to go hand in hand,” he wrote. “Yet designers often tend to discard the balance between design and function, creating gorgeous data visualisations which fail to serve their main purpose – to communicate information.”

Market researchers make similar mistakes, but for them the balance skews the other way. If designers are sometimes guilty of prioritising what something looks like over its ability to convey a story, researchers are often guilty of trying to cram too much information into visualisations to the detriment of the narrative.

“Most other businesses are very organised and have a thorough understanding of how you present yourself and how you present data, and how important it is, but it struck me that the research industry generally didn’t,” says Martin Lambie-Nairn, founder of Lambie-Nairn & Company and the man responsible for the original Channel 4 logo and the 1997 rebrand of the BBC.

In July last year Lambie-Nairn was hired by research group TNS to become its first creative director. He saw the job as a “challenge”, he said. “Essentially, you’re asking an organisation – and the industry generally – to have an appreciation of a particular process that has not been traditional to them. What we had to do was create a [visual] language, show how it worked and change people from the inside.”

CycleFlows

A still from an animation produced by City University professor Jo Wood showing the first five million journeys taken using London’s public bicycle hire scheme. The three clusters are (from left to right) Hyde Park, King’s Cross/St Pancras and journeys to and from Waterloo Station to the City

Drawn to scale

Visualising data is not new. Line charts, pie charts and bar graphs are traditional ways of telling a story, based on numbers, in visual form. But data visualisation tends to imply something more… impressive. For instance, network maps of the blogosphere, renderings of the social media connections between people in different parts of the world or David McCandless’s Billion Pound-o-Gram comparing different aspects of public expenditure.

Each of these has something in common. They are ways of dealing with huge numbers and masses of data points and presenting them in a way that makes it easy for people to comprehend things that would otherwise be difficult to think about without a visual aid.

City University’s Jo Wood is the man behind a popular recent visualisation showing the journeys taken by users of London’s “Boris bikes”, the public bike hire scheme. Wood is professor of visual analytics, a discipline that has emerged out of the US home security and defence sectors. “The reason it’s arisen in those sectors is that they have the challenge of huge volumes of data and they need to pick out particular patterns – whether it’s terrorist activities or unusual behaviour or whatever,” says Wood.

Essentially, visual analytics is designed to address a class of problems that says, “I’ve got a massive dataset. Buried in there somewhere is something really important and I need to find it,” Wood says. “Traditionally that sort of analysis has been done through data mining and the like. But visual analytics has emerged because there was a realisation that you need much more human input into that process of spotting the unusual or the important.

“We’re dealing with datasets that are much more complex and much larger than they ever used to be. There are so many ways in which you could potentially explore a dataset that you need some kind of guidance… We need this visual navigation to tell us where to look or what to look for.”

David McCandless, a journalist turned information designer, agrees that, “Using a visual mechanism makes it easier to spot trends, apprehend patterns and see outliers. I imagine very experienced researchers can see that kind of stuff in spreadsheets,” he says, “but the fact that it’s easier visually makes it more widely accessible.

“If you’re interested in telling a story, then obviously a visual language has a broader appeal: it draws the eye, is attractive and impactful and cuts through the blizzard of information we’re often navigating through each day.”

McCandless graphic

Stock Check, by David McCandless’s Information is Beautiful Studio, shows the estimated remaining supplies of non-renewable resources and how many years until they run out. Smartphone owners beware – indium-dependent touchscreens seem to have a very limited shelf-life

Provoking a reaction

For people like Stan Sthanunathan, vice president of marketing strategy and insights at The Coca-Cola Company, impact is everything. “What we want information to do is inspire and provoke people.” Data visualisation is just one way to achieve this, he says. Sthanunathan thinks researchers should think about storytelling “a little more broadly”, but there are very good reasons why clients are wanting more from their agencies’ deliverables.

“People are getting increasingly tired of busy charts,” he says. “People don’t have time to look at them.” It’s not only researchers that are guilty of this, he says, but “most people get confused between a fact-based presentation and a fact-filled presentation”.

“You have to go through all the due diligence to figure out what the story is,” says Sthanunathan. “But your story will only require some of the data that you have in your study. When you’ve chosen the data that you want to use, spend some time thinking about how to bring it to life.” Within Coca-Cola, Sthanunathan says the process is to identify two or three “killer facts” that the insight team wants the senior executive team to remember. Within General Mills, meanwhile, the focus is on coming up with “the killer chart”, says Jeffrey Hunter, a former consumer insights director at the firm, now a consultant. “This is the chart that makes an argument in a unique, clear and compelling way.”

Hunter says: “I generally think of data visualisation as something we do with very complex or big data. General Mills abhors complexity (and jargon) and data visualisation of complex, big data or statistical outputs would of necessity be reduced to something very essential.” Suppliers won’t be surprised to hear that time is of the essence. Asked whether he’d be prepared to wait for his agency to deliver a piece of data visualisation, Sthanunathan says: “Fancy graphics are nice but if it’s going to take an extra week then I would prefer to just have the data earlier.”

For big companies like TNS, then, the focus is on “industrialising” the data visualisation process. “We have built a system that works in such a way that it allows people to produce PowerPoint decks in a matter of seconds,” says Roeland Nieuwenhuis, a senior consultant who has been working with Martin Lambie-Nairn on what the agency calls Project Everest – a name that speaks of the scale of the task facing researchers, whether technical or cultural.

“The automatic charting outputs are really quite wonderful,” says Aziz Cami, creative director for Kantar, TNS’s parent company. “They’re bespoke pieces of work which are linked to databases. Of course, the outputs at the moment are fundamentally PowerPoint-based, but we’re also looking at more web-based opportunities because we feel that’s where the future lies.”

stocktrade

These images from Stamen’s Eric Rodenbeck show a minute’s worth of trading on the Nasdaq exchange. In the top image, a single trader (in lime green) is responsible for the majority of shares. Below that, the yellow dots in a row show trades in a single stock at a single price

Time to play

Everybody interviewed for this supplement imagined a world where data visualisations become much more interactive and ‘live’ – with real-time information feeding in to them. How soon we get there, though, is a question of technology. Despite the availability of a host of free and paid for tools these sorts of interactive portals still take “a hell of a lot of time and money to put together,” says Ian Jarvis, who leads the data visualisation working party at Ipsos Mori.

“At the moment if we do these, even with the reporting portals we’ve got, it generally takes a month or so to set them up because you’ve got massive datasets that have to be formatted and organised properly and run through to make sure everything’s perfect. Then you’ve got to design the portal and make sure the data links in properly.” Meanwhile Jarvis questions how much freedom people should get to work with real-time or massive datasets. “The danger is if you open up entire datasets to people they can do their own analysis and get it completely wrong.”

And yet giving people the freedom to play with data is important in helping them get comfortable with it. Eric Rodenbeck, founder of design studio Stamen, says the aim of his work is “to entertain by informing and inform by entertaining”. Much of what he does is for public consumption – “I’m not super-interested in tools for CEOs,” he says. “I definitely want people to engage with our work and learn something – I just want them to have fun first.”

Rodenbeck’s belief in the “fun” potential of data visualisations positions him at the opposite end of the scale from the research buyers and suppliers who are looking for more in the way of function. As such, he rejects the idea that viewers need to be able to understand everything a visualisation is trying to tell them right away. “The whole point of doing this work is to make something that’s visually robust or visually seductive enough to engage people initially, but which gives them more the more time they spend with it. It’s not something that reveals itself all at once.

“I don’t want to go on record as being intentionally obscure, but I do think there’s a lot of room in this space for work that is richly interactive, not just immediately apparent. There’s so much that you can learn by using your eyes. For me, data visualisation is not just about answering the questions you already have, it’s about finding better ways to ask new questions.”

2 Comments

12 years ago

Unquestionably new data visualisation techniques have brought data to life in wonderful ways. Not least because they have translated the 'beauty' of numerical relationships into visually engaging pictures and films. But from a market research perspective have we really learnt anything we didn't know? Have we generated dramatic insights that change the decisions we make? I would argue not - we still earn our keep explaining what the numbers mean rather than finding pretty ways to present them. A nice change from Power Point though but.

Like Report

12 years ago

Thanks for this timely reminder of the importance of using the right tools for our story-telling. Who, if not the MR industry, should be leading the field in data visualisation? We're the experts in telling stories with data! Unfortunately, it's hard to get people to actually think about the tool they use when PowerPoint has been ingrained for so long.

Like Report