The little of visualisation design: Part 65

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 subtle use of coloured apparatus to indicate categorical association. The graphic in this case was produced by Gus Wezerek for an article titled ‘Which 2020 Candidates Have The Most In Common … On Twitter?‘. Using a matrix of Venn diagrams, this explores the proportions of twitter followers that are common to each Democratic candidate pairing.

This work alone could provide a year’s worth of little observation but I’m just going to focus on the smart way colours are used along the apparatus of the matrix to inform the reader which Venn diagram set relates to which candidate. You might think that to show this many distinct combinations would involve lots of varied categorical colours, but though there are ten candidates, it is the ‘two’ that matters in each diagram case – whose circle set is on the left and whose is on the right.

The little of visualisation design: Part 64

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 irregular axis scales. The piece of work in focus was produced by Allison McCartney of Bloomberg in an article titled ‘Ford Explorer Owners Say Their SUVs Are Making Them Sick‘.

The little design choice I like about this chart is how only the relevant y-axis axis scales are displayed. Think ‘why we include such apparatus?’ – we do it to assist the reader in judging the size (or technically, relative position) of the point marks in the chart. Given the distribution of the carbon monoxide readings plotted along the ‘parts per million’ y-axis, it therefore makes sense to provide axis-scale intervals of size 10 for the range 0-60, where most items are concentrated. However, with the remaining five items being spread more widely from 60-140, the scale switches to intervals of 20.

The little of visualisation design: Part 63

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 step-by-step annotations to provide reading assistance. The work in question was created by Matthew Smith for YouGov about the dominance of Tesco supermarket.

What I like about this work (and tip o’ the hat to Robert Grant for alerting me to it) exists in two layers of help. Firstly, the viewer is guided through the task of perceiving the heatmap, with numbered captions provided to explain what the values down the columns mean….

… and what the values across the rows mean.

Secondly, a caption explains the meaning of an single intersecting cell, using plain English to help support the viewer’s interpretation rather than leave them to spend time forging their own understanding. This is a great example of a designer caring about an audience and coaching them through the reading process.

The little of visualisation design: Part 62

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 potential consequence of misinterpretation caused by small (and probably very innocent) graphical errors. The work in question was published by the BBC during the recent heavy snow to offer advice about ‘stopping distances in snow and ice‘.

The problem with the original graphic, shown above, is the way a viewer may use the small car images to perceptually orientate their judgment of the size of the bars on the right hand side, showing the thinking and braking distances required for different speeds in different weather conditions. At first glance you might think that when travelling at 30 mph, in normal conditions, you only need to leave about a half-a-car length of distance for stopping, when the actual 23 metre stopping distance would be closer in equivalence to around 4/5 standard car lengths. (There is also confusion through the colouring of the cars and misleading association with red = thinking distance, yellow = braking distance).

The original graphic has now been modified, probably due to the challenging coverage from commentators like Ben Goldacre on Twitter. Now the car icon is removed and only the bar measurements remain. Though this is an improvement, personally I would find it easier in a driving situation to think in terms of car lengths as the unit of judgement rather than only distances.

The little of visualisation design: Part 61

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 composition choices that exploit shapes and spaces formed by charts, text and images. The work that demonstrates this idea comes from a National Geographic project titled ‘When the too-early bird sings‘, created by Monica Serrano and Ryan T Williams.

What I like in this piece is how the semi-circular chart layouts create a small shape of emptiness that perfectly accommodates the associated bird images. In the main chart, for the robin, notice how the placement of the dawn/dusk icons and descriptive text continue to create a nice archway-like balance and alignment between all elements. Furthermore, and most noticeable with the smaller graphics down the right-hand column, the layout of the left-aligned bird titles and descriptive text below each chart uses the available space so efficiently preserving harmony with the chart and image above.

The little of visualisation design: Part 60

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 subtle styling of value labels. The work in focus is one of my favourite (ever) visualisation pieces titled ‘Simulated Dendrochronology of U.S. Immigration: 1790-2016‘ and developed by Pedro Cruz, John Wihbey, Avni Ghael, and Felipe Shibuya.

Most who have seen this before will be familiar with the main tree ring view of immigration across the whole of the U.S., but I want to focus on a further variation in this project, assembled by Steve Costa, showing a cartogram of tree rings for all 50 states. One of the little design features I loved about this was the presentation of the state labels in the way they are curved around outer edge of their associated tree ring shape.

This simple choice seems to create a glue effect, binding the label to its relevant tree ring and eliminating any perceptual doubt about which relates to which that could have been the case with straight labels. It creates more cohesion between the representation and the annotation, rather than making the latter feel like an intrusive, administrative after-thought.

The little of visualisation design: Part 59

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 careful balanced design of value labels. The piece in focus was produced by Ashley Kirk for The Telegraph in an article about police response times (preview here if you’re not registered).

The heat map reveals patterns of average response times to emergency calls for police forces across England and Wales over several years. The colour scheme shows the darker shades edging towards or exceeding target.

What particularly struck me when I first looked at this piece was an appreciation for the careful thought that would have gone into seeking the right balance in the visibility/prominence of the value labels within the chart, especially in contrast to the cell colour encoding. Designing the presentation of text annotations – titles and intros, captions, axis labels, value labels – is an intricate matter involving typeface, font style, font size and colour decisions.

By making the value labels white (and possibly with some slight opacity, though that could be an illusion), you can see the labels when you want to see them but they don’t shout. Or, more to the fact, only the important ones are shouting – the values within the darker cells that represent perhaps the most important values of longer response times. A grey shade couldn’t be used as it was already committed to showing no data or no target, and black would have created more contrast for the least important values.

The little of visualisation design: Part 58

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 coloured annotations to highlight key patterns in a chart’s display. The piece in focus was produced by Tim Meko for the Washington Post in an article ‘Drenched city: 2018 is now Washington’s wettest year ever recorded‘.

The chart shows patterns of cumulative rainfall in Washington DC with a separate line for each year going back to 1871 when records began. You see the top-most line representing the previous record amount of annual rainfall set in 1889, with 2018 now surpassing that total.

What I particularly liked in this chart were the yellow regions that draw attention to the intense periods of downpours during May and July/August. Sometimes, when you see a line slope you can lose a little orientation about the true magnitude it has grown by. These yellow snapshots emphasise just how much the total has shot up vertically within such a short period of time horizontally.

The little of visualisation design: Part 57

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 often-discussed issue of a non-zero baseline. The work in question is a chart found in this article published on the BBC News website about the decline in the value of the pound vs. the dollar.

There is lots of debate about whether a line chart needs to start from a zero baseline position. Unlike the bar chart, for which the encoding is based on the perceived size of a ‘bar’ line, the line chart is based on a series of point marks along a quantitative scale joined by connecting line to reveal localised trends. It is the perceived slope of these connecting lines that really stirs the debate because any change in the y-axis, and indeed x-axis, scales alters the apparent steepness or flatness of the trend lines. My approach is, boringly, to take each chart on a case-by-case basis and focus on ensuring no confusion is created in the eyes and mind of the audience. More broadly, I tend to only commence from zero when zero means something significant to that subject.

Anyway, this post isn’t about covering that debate as such as pointing out how the simple use of a thicker, prominent stroke across the x-axis baseline can inadvertently imply that the associated y-axis value is zero. As you see in the chart above, the lowest value range is 1.24 not zero which can alter the interpretation of the relative downturn in the pounds fortune if you don’t pay close attention. I would have presented the baseline using the same style as the other gridlines, a thinner grey line, and maybe added a bit more empty space below to avoid any potential confusion (good further proposal here)./p>

The little of visualisation design: Part 56

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 dissonance of seeing a negative value plotted in a way that doesn’t immediately convey negativity. It relates to a chart found in this article published on The Guardian about the demise in sales of DVDs and other products reaching the end of their life, at least in the eye of the retailer John Lewis.

The chart above presents seven products sold by John Lewis that were in significant decline sales wise during 2018, showing their percentage decrease over 2017. The way the chart is presented, with bars going to the right and the axis labels showing positive values, is confusing given the values represented should be negative. Quite simply, I would have reversed the chart, with the labels on the right and negative bars growing to the left along a negative axis.

The Guardian know what they’re doing and this stands out as a blip. In trying to work out what might have led to this, I have a suspicion it might be because of some charting template limitations. The inclusion of the negative value labels in the category headers indicates to me an attempt to rescue things, but its rather too late.

A second chart showing those products that have increased in sales during 2018, compared to 2017, merely compounds the confusion as it presents the (now positive data values) using the same chart design as the first. Though I feel this is the right way to convey this particular data, I might have considered using colour to further differentiate this chart’s positive values from the negative ones in the earlier one.