Changing the meaning of 'a good data visualisation'

Everyone loves a good data visualisation, and they’ve never been more in-your-face than in the last couple of years. Whether  mapping local language variations  or the way the wind is blowing there's something new every day to whet your appetite for pretty data.

These visualisations are all driven by a widespread idea: that computers can do amazing things with data that could never be done by people using paper. Computers can make data move and pulse and zoom and scroll. Data on paper just sits there - immobile and boring.

But I don't think that zooming and swooshing are the most important thing that computers bring to data visualisation.

In fact, I think the most  important advantage that computers have over paper is the ability to hide unnecessary information. 

The rest of this post is my attempt to persuade you why this matters.

Why hiding information is so important

Let me explain by critiquing that most hallowed of all data visualisations: the Tube map. Here it is, in the current 2012 official printable form:

Over the last decade there have been dozens of riffs and variants* on this map: a new generation of design and data visualisation geeks tipping their hats to the genius of Harry Beck.

However, I think this reverence blinds us to something that should be staring everyone in the face: for an average user of the Tube the famous map contains approximately 100 irrelevent pieces of data for every one piece that actually matters to them. The data contained within the map is literally about 99% irrelevent to a user trying to get from A to B, especially for the 26 million tourists who allegedly visit the city every year.

This isn't Beck's fault. He did a world-changing job with the tools he had: paper, pens, typography and a brilliant design sense. 

For 100 years this was the way things had to be. But then the web came along, and it suddenly became possible to show people information about their journey that was 70% or 90% relevent - not 1%. How? Through the data visualisation known as the Transport for London journey planner. Here's a sample journey:

Now I'm not going to defend TfL's web design here: obviously whoever designed it is no modern Beck. But my point is that  the proportion of this screenshot which contains user-relevent information is hugely higher than can be contained on any map, even one designed by a legend.

This is because the computer makes it possible to conceal everything the user doesn't need to know: it takes the tiny sub-section of the Tube network relevent to this passenger and turns it into instructions, with the key part of the data visualisation being a scripted sentence. Data is visualised in the form of words, and a few icons. 

This is a data visualisation that uses the capabilities of the technology to give the users what they need. It deserves greater respect than most sliding, zooming, pulsating online data visualisations you will see.

The Moral

I want to see more data visualisions that make it clear that their authors have really thought about why the computers they have in front of them can help people more effectively than a quality paper design could. The paper Tube map is still more fundamentally useful than 99% of hot web visualisations, even though it was crafted on an infinitely more limited technology. Why aren't more visualisations simply better, given the power at our fingertips?

I think it comes down to the way that data visualisation gets increasingly framed as an art. I have no problem with people using computers to generate art, but I think that data visualisers are not best seen as artists: they are better seen as craftsmen building tools to help people to make their way through life. We should aspire to be Stradivarius, not Picasso.

 

* Whilst there are dozens of redesigns of the almost-perfect Beck map, there are virtually none of the obviously-flawed TfL journey planner. I think it is worth pondering why so many people want to tinker with the almost unimprovable paper visualisation, whilst ignoring the pressing design problem with a fundamentally more user-centric visualisation that gets 3m visits a month.

Comments

[Robinhouston]

Presumably this post is intended to expand on the line of thought you explained your previous post on The Two Types of Data Visualisation, and to advocate ‘answer visualisations’ over ‘story visualisations’ using the contrast between the Tube Map and TfL’s Journey Planner as an illustrative example.

The question what makes a data visualisation ‘good’ is an interesting one, and since goodness may be found in many forms perhaps it is helpful to differentiate varieties of goodness. The two I propose for consideration here are ‘useful’ and ‘interesting’.

This relates to your division into story visualisations and answer visualisations: stories are supposed to be interesting, whereas answers are supposed to be useful.

It seems to me that the Journey Planner is both more useful and less interesting than the Tube Map. More generally I think there’s a tension between the kind of stripped-down utility you’re advocating in this post, that may indeed be well-served by excising all details not germane to the task at hand, and the fascinating digressive quality of a well-designed bird’s-eye view of some complex collection of information.

The US wind map is of no use to me whatsoever, but it is delightful and fascinating. Watching the flow of air across the surface of the US raises so many questions. What are the topographic factors influencing the flow? Where are the mountains and valleys? What is the distribution of air pressure, and why? Is today typical or atypical? Where are the prevailing winds? It doesn’t provide any answers – I didn’t start with any questions that such a map would be capable of answering – but, oh, the questions! That is what makes something interesting, that it inspires questions, opening up the space of ideas rather than closing it down to a single answer.

The glory of ‘What size am I?’, of course, is that it is both useful *and* interesting. Though as it happens it is of no use to at all to me myself, because I am not a woman and I do not often buy clothes, but I love it because it exposes to me an aspect of the world of which I had previously been altogether ignorant.

If this is really what it comes down to, are we eager to stand for usefulness at the expense of interestingness? I don’t think I am.