The little of visualisation design: Part 52

This is part of a series of posts about the ‘little of visualisation design’, respecting the small decisions that make a big difference towards the good and bad of this discipline. In each post I’m going to focus on just one small matter – a singular good or bad design choice – as demonstrated by a sample project. Each project may have many effective and ineffective aspects, but I’m just commenting on one.


The ‘little’ of this next design concerns using containers to separate different scaled views of analysis that would otherwise be distorted using a single view. The visualisation in question is from a Washington Post in an article about ‘The Stock Market Decline May Not Be Over Yet‘.

As you see in the graphic, there are three distinct containers using slope graphs to show stories of how the 26 occasions when a daily +4.8% drop in the Dow has occurred. The first view shows these 26 daily drops, the second view shows the status of the price in each case one month later, the final view shows the price one year later.

Rather than trying to somehow squeeze these distinct perspectives into a single chart, each is given separate space to breath. Using separate containers to give a home to these different frames of data is a simple tactic that helps build the narrative and preserve clarity.

The little of visualisation design: Part 51

This is part of a series of posts about the ‘little of visualisation design’, respecting the small decisions that make a big difference towards the good and bad of this discipline. In each post I’m going to focus on just one small matter – a singular good or bad design choice – as demonstrated by a sample project. Each project may have many effective and ineffective aspects, but I’m just commenting on one.


The ‘little’ of this next design concerns, ironically, a method of showing how big something is by showing it exceeding the available screen space. The visualisation comes from The Guardian in an article about ‘Bezos’s empire: how Amazon became the world’s most valuable retailer‘.

The basis of the visual is a vertical scroller that steps you through Amazon’s annual revenues generated yearly since 1996. As you scroll down the page, the bubbles grow proportionately with the growing revenue, starting off with the smallest of $15.7m. Year by year the bubbles grow until a critical ‘visual’ point is reached around 2013 and through to the most recent value at 2017. As you see below, the value size is now such that the bubble size transcends the available vertical space of the chart canvas and screen dimensions.

I’m viewing this page on a laptop but, as you see below, the same effect is seen on the mobile version, now with horizontal space restrictions from the portrait view in contrast to the vertical limits of a landscape laptop window. What we’re seeing here is a subtle but clever narrative device that essentially says ‘this value is now so ridiculously large that you no longer need to concern yourself with trying to perceive its size, just see it as being big, so big it fails to fit in your screen’. The value label is there if you want to know the actual magnitude but otherwise ‘the gist’ of these latter numbers being extremely large is sufficient for the point being made.

The little of visualisation design: Part 50

This is part of a series of posts about the ‘little of visualisation design’, respecting the small decisions that make a big difference towards the good and bad of this discipline. In each post I’m going to focus on just one small matter – a singular good or bad design choice – as demonstrated by a sample project. Each project may have many effective and ineffective aspects, but I’m just commenting on one.


The ‘little’ of this next design comments on the unnecessary burden we can put on our readers’ brains to read what should otherwise be simple charts. The graphic in question is a pictogram taken from an article on the BBC titled ‘Brexit: What’s happening to migration?‘, illustrating the UK’s net migration figures in the 12 months before and after the Brexit vote.

There are two perceptual obstacles with the design of this chart. Firstly, the small icons representing units of 10,000 people works fine for the first value but not for the second, where the final icon has to be sliced up to 60% of its width to represent a unit of 6,000 people to make the final total count of 336,000.

The second issue relates to the composition of the chart and how it uses row blocks of up to 15 icons. As readers we can’t just fall back on our natural counting heuristic, grouping ‘chunks’ of units into 10s to quickly arrive at a total count, instead we have to count out the icons one-by-one to learn that there are up to 15 per row. This takes far too much time to make it worth the effort. You might argue the value labels tell us the values and the chart just offers a quick visual portrayal but in offering the units the eye is inevitably lured in to counting. It could have been so much quicker and easier using two sized-bars or shapes.

New project: The Malofiej Experience

Every year, for a few days during March, visual journalists, designers and developers from across the globe assemble in the city of Pamplona. The University of Navarra is the venue for the Malofiej Awards and World Summit, the annual celebration of the very best in infographic and data visualisation design.

In 2012, I was invited to the 20th edition of this prestigious event. Just attending Malofiej was a privilege in itself, but to participate as a judge and speaker alongside some of the most talented people in the field was a defining experience.

2017 was the first opportunity I had to return to Pamplona. As this was the 25th anniversary event, I wanted to mark this special occasion and salute its significance to me and my career.

Despite having no experience, no knowledge and, frankly, no talent, I aspired to make a short film to document the essence of what makes Malofiej such a unique and enduring event.

Enlisting my trusted co-producer, Matt Knott, armed with only two iPhones, two microphones and two stands, we jetted off from the UK bound for Pamplona to interview some of the principal protagonists and to capture THE MALOFIEJ EXPERIENCE.



Thank you to all interviewees who generously offered their time and contributed such astute observations to this film. Though we couldn’t accommodate footage from all contributors, every interview was hugely valued and appreciated. Below are the names and organisation credits of those who appear in the film, in order of appearance.

Anna Flagg, The Marshall Project

Javier Zarracina, Vox

Karin Schwandt, Schwandt Information Design

Andre Gouws, Media24

Jen Christiansen, Scientific American

Sarah Slobin, ThomsonReuters

VY Chow, Apple

Kennedy Elliott, National Geographic

Maarten Lambrechts, Freelance Data Journalist and Visualisation Consultant

Jon Schwabish, Urban Institute

Nathaniel Lash, Tampa Bay Times

Michael Brenner, Beyond Words Studio

Nigel Hawtin, Nigel Hawtin Information Design

Liz Gould, Cancer Research UK

Isabel Meirelles, OCAD University

Kaitlin Yarnall, National Geographic

Hiram Henriquez, H2H Graphics & Design Inc

Simon Scarr, ThomsonReuters

Nicolas Vargas, Corriere della Sera

Fernando G. Baptista, National Geographic

Emilio Amade, El Mundo

Tom Giratikanon, New York Times

Manuel Canales, National Geographic

John Grimwade, John Grimwade Information Graphics

Frédérik Ruys, Vizualism

Finally, special thanks to Javier Errea for indulging our request to produce this film without hesitation or restriction, and for granting us the rare privilege of a premier showing of this film at Malofiej 26.

The little of visualisation design: Part 49

This is part of a series of posts about the ‘little of visualisation design’, respecting the small decisions that make a big difference towards the good and bad of this discipline. In each post I’m going to focus on just one small matter – a singular good or bad design choice – as demonstrated by a sample project. Each project may have many effective and ineffective aspects, but I’m just commenting on one.


The ‘little’ of this next design concerns the use of visual aids to help establish a sense of scale. The graphic in question is taken from a piece in The Economist titled ‘The Falcon Heavy’s successful flight is another vindication for Elon Musk‘. For those without a subscription, you can see the image posted on the Head of Graphics, Phil Kenny’s, timeline.

This graphic illustrates a selection of launchers (from NASA, NPO Energia and SpaceX), in light of the successful Falcon Heavy launch, to provide comparisons of size, payload and periods of operation. What I like about this piece is the subtle inclusion of the Statue of Liberty silhouette in the empty background space on the left of the illustrations. This helps provide the reader with some extra sense of scale to appreciate the size of the rockets displayed in the chart, supplementing the units of size plotted on the y-axis.

That isn’t the only useful visual comparison, as Kristine Briones astutely observed – right at the base of the chart you will find tiny human figures to the far left and far right of the x-axis. Lovely work by The Economist’s graphics team.

The little of visualisation design: Part 48

This is part of a series of posts about the ‘little of visualisation design’, respecting the small decisions that make a big difference towards the good and bad of this discipline. In each post I’m going to focus on just one small matter – a singular good or bad design choice – as demonstrated by a sample project. Each project may have many effective and ineffective aspects, but I’m just commenting on one.


The ‘little’ of this next design concerns the choice of axis scale intervals. The graphic in focus is taken from a screen shot of BT Sports (UK) TV coverage of the Australia vs. England 1-day cricket match. I suspect the visuals are produced by the host broadcasters in Australia (most likely Channel 9).

This chart shows the cumulative progress of run scoring for each team across the 50-overs that each team bats for. The chart helps to show whether Australia appear to be on track for victory or otherwise given England’s rate of run scoring. My issue with this chart’s design concerns the y-axis scale intervals that use units of 70 runs.

It isn’t that you CAN’T get a sense of the quantitative positions of the line heights but more that it puts greater burden on the eye and the brain to work out what the values are between these intervals. Surely it would be much easier to perceive if the chart used axis intervals of 50, 100, 150 etc.? Judging runs in units of centuries and half-centuries is integral to the language, focus and statistical chunking of cricket so I just do not see the merit of using the 70, 140, 210 scales.

The little of visualisation design: Part 47

This is part of a series of posts about the ‘little of visualisation design’, respecting the small decisions that make a big difference towards the good and bad of this discipline. In each post I’m going to focus on just one small matter – a singular good or bad design choice – as demonstrated by a sample project. Each project may have many effective and ineffective aspects, but I’m just commenting on one.


The ‘little’ of this next design concerns the creative placement of chart labels across a panel of small multiples. The piece in question comes from Zachary Labe looking at the last 70-years of autumn temperature anomalies in the Arctic.

One of the features I really liked about this piece is the clever way Zachary incorporated the chart labels for each yearly slice. Rather than positioning the year centrally aligned above the circular panels, which might be the default approach, he astutely slotted them in the vacant space just above and to the right of each circle and slightly rotated at a 45° angle. Their adjacency is preserved but, crucially, the overall grid layout is nicely compacted with no wasted empty space required to accommodate the seven rows of labels.

The little of visualisation design: Part 46

This is part of a series of posts about the ‘little of visualisation design’, respecting the small decisions that make a big difference towards the good and bad of this discipline. In each post I’m going to focus on just one small matter – a singular good or bad design choice – as demonstrated by a sample project. Each project may have many effective and ineffective aspects, but I’m just commenting on one.


The ‘little’ of this next design concerns potential redundancy in axis labelling. The focus of this piece is a project published by The Pudding, titled ‘The Office Dialogue in Five Charts‘. Unlike previous posts in this series, this piece is not taking a critical perspective, rather I’m exploring some alternative approaches to offer a reminder of the subtleties of choice that exist about seemingly low-level and mundane matters. I really like the project in focus but I’m wondering to what degree the repetition of chart annotation is useful?
(Note: You might need to click on the images to see them in higher resolution)

In this series of small multiple area charts we see y-axis interval labels 0% >> 30%, x-axis interval labels 1 >> 8 and the x-axis title ‘Season’ repeated across every one of the 20 charts. This unquestionably aids the direct readability of each panel but much of it could arguably be described as redundant. If we assume that there were no programmatic restrictions (in some situations you might not be able to customise the design of individual charts in a matrix like this) one could just have the first panel in each row showing the y-axis intervals and the first chart only titling the x-axis…

You might need to keep in the x-axis intervals because they don’t have vertical gridlines to aid reading downwards from the first row. As you see below, stripping x-axis interval labels from rows two, three and four makes them a little harder to read. Maybe this is too severe. However, it might also enable the eye to focus less on apparatus and more on observing the patterns in the data – placing more emphasis on ‘the gist’.

As ever the boring tag ‘it depends’ carries as much weight as ever here as does the notion of the goldilocks principle for annotation design – the sweet spot existing between not too much and not too little.

The little of visualisation design: Part 45

This is part of a series of posts about the ‘little of visualisation design’, respecting the small decisions that make a big difference towards the good and bad of this discipline. In each post I’m going to focus on just one small matter – a singular good or bad design choice – as demonstrated by a sample project. Each project may have many effective and ineffective aspects, but I’m just commenting on one.


The ‘little’ of this next design concerns a neat way of squeezing more value from the space afforded to a project heading/title. The demonstration of this technique was found in a piece by Cath Sleeman for Nesta, titled ‘The gender imbalance in UK film casts‘.

This is a very straightforward device but something I’d never considered nor seen before. In the snapshot image above you see that the title space includes background colouring that doubles up as a stacked bar chart, showing the headline % split between male and female on-screen roles in movies. By placing this single, simple chart in the relatively dead space that would otherwise exist beneath and around the heading, it efficiently offers up an immediate visual to support the focus of the report that follows.

The little of visualisation design: Part 43

This is part of a series of posts about the ‘little of visualisation design’, respecting the small decisions that make a big difference towards the good and bad of this discipline. In each post I’m going to focus on just one small matter – a singular good or bad design choice – as demonstrated by a sample project. Each project may have many effective and ineffective aspects, but I’m just commenting on one.


The ‘little’ of this next design concerns z-sorting and how you manage the depth-arrangement of overlapping value series. The chart in question comes from an article by FiveThirtyEight titled ‘The National Parks Have Never Been More Popular‘.

The bump chart plots the relative popularity, in ranking terms, of the US national parks going back over 100 years. With an increasing range of recognised parks and, one must assume, an ever-expanding register of visitor data, there is a lot to plot in one chart and a risk of creating lots of noise and limited signal.

Three key things happen here. Firstly, only 11 of the parks are coloured, creating editorial emphasis versus all the other categorical series, which are included but presented in grey and without labels. Secondly, you will see there is a white outer-stroke wrapped around the coloured lines to help them pop-out a little further. Thirdly, and the main thrust of this post, is the decision to apply z-sorting on the coloured lines whereby the most recent rankings influences the overlapping order when lines meet back through the timeline. So, for example, The Great Smoky Mountains line is always on top, the Grand Canyon will then be on top of all other lines apart from the Great Smoky Mountains etc.