6. Consider these numbers…
Property Value
Mean of x 9
Variance of x 11
Mean of y 7.5 (to 2 dp)
Variance of y 4.122 (to 3 dp)
Correlation 0.816 (3dp)
Linear regression y=3+.5x (3dp)
27. “When I am working on a
problem, I never think
about beauty. I only
think about how to solve
the problem. But when I
have finished, if the
solution is not
beautiful, I know it is
wrong.”
- R. Buckminster
Fuller
Slide 5: viz of Anscombe’s quartet
But what if you VISUALISE them: wow! They are clearly NOT the same. This is Anscombe’s quarter – a set of numbers designed to show the importance of seeing your data. Your brain processes this info inistantly: no analysis necessary. Imagine if you were getting this info from your data.
Slide 5: viz of Anscombe’s quartet
But what if you VISUALISE them: wow! They are clearly NOT the same. This is Anscombe’s quarter – a set of numbers designed to show the importance of seeing your data. Your brain processes this info inistantly: no analysis necessary. Imagine if you were getting this info from your data.
Slide 5: viz of Anscombe’s quartet
But what if you VISUALISE them: wow! They are clearly NOT the same. This is Anscombe’s quarter – a set of numbers designed to show the importance of seeing your data. Your brain processes this info inistantly: no analysis necessary. Imagine if you were getting this info from your data.
Slide 5: viz of Anscombe’s quartet
But what if you VISUALISE them: wow! They are clearly NOT the same. This is Anscombe’s quarter – a set of numbers designed to show the importance of seeing your data. Your brain processes this info inistantly: no analysis necessary. Imagine if you were getting this info from your data.
Slide 5: viz of Anscombe’s quartet
But what if you VISUALISE them: wow! They are clearly NOT the same. This is Anscombe’s quarter – a set of numbers designed to show the importance of seeing your data. Your brain processes this info inistantly: no analysis necessary. Imagine if you were getting this info from your data.
Slide of the amazing eye! Or maybe the monkey to man to chart
Back when we were knucle draggers, it was important to find the right food, avoid the poisonous berries and not get eaten by predators. For that, we evolved shit hot eyes. Little did those neandarthals realise that the hard work of evolution they were putting in would be just right for looking at data.
And our eyes are AMAZING! We can see and interpret things instantly, if we play to our strengths.
Is it easy to answer the question on this slide?
Now is it easy? Yes it is – we used a preattentive attribute, colour, to “pop out” the 9s. Our brain starts processing differences in colour before we begin to make conscious understanding of the data
Length and colour are two preattentive attributes: things we perceive BEFORE we consciously address them with our mind.
Which is bigger? A, B or C?
Which is bigger? A, B or C?
And that’s why certain charts are better than others for PERCEPTION. The bar chart may be “boring” according to some people, but boy, it’s the best way to see differences in values.
Which is bigger? A, B or C?
Right – here we get into a debate. When you’re designing charts, how do you balance functionality with aesthetic design?>
When designing visualisations for others, you need to balance the ability for the them to actually understand what the heck you’re trying to say with a design that makes them want to look at your chart.
There are so many right and wrong ways to do this it is not possible to tell you what’s right: it depends on your goal. On your audience. On the data. On the questions.
At one end of the spectrum you have something that’s built for aesthetic value – the Colour wheel is a good example. It looks engaging but if you try and gain insight from it, it’s hopeless.
At the other end is a purely functional business dashboard. It’s very austere, aesthetically, but it is designed exactly to play to the strengths of the visual system
Slides of pie chart and complex legend or crazy 3d invented chart (chord diagram? Radial bar chart? McCandless cover with the colours?
the visual system is flawed. Change blindness and the Pepsi/Aubergine Just do one.
This is important because charts need to be designed to avoid making the brain do more work than necessary. Everytime you change things, or force your audience to move away, write code, switch view, they are flushing their cache in the brain and it’s costing them time.
However look at this. This is an award winning viz made by Simon Scarr around the time the US announced its withdrawal from Iraq. It is designed using visual best practises we have discussed. It is hugely compelling at the same time as being designed to draw you in. It’s as far as you can get from MccCandless as possible. But better.
Right – here we get into a debate. When you’re designing charts, how do you balance functionality with aesthetic design?>
When designing visualisations for others, you need to balance the ability for the them to actually understand what the heck you’re trying to say with a design that makes them want to look at your chart.
There are so many right and wrong ways to do this it is not possible to tell you what’s right: it depends on your goal. On your audience. On the data. On the questions.
Finally, even when sticking with good design principles you have loads of options to consider when designing. And these choices will control the message your viz has. Every viz should have a message. Let’s see what we can do with this.
First off, let’s look at 3 things he did
The title: that’s enough to tell you his message: he’s emphasising the LOSS of life.
Was his choice of red neutral? Of course not – it’s blood.
Bar charts normally go up. What does this one go down? To look like a blood smear.
What can we do to this chart to change the message without even changing the vizzes>
Rotate the viz. Now the drop in deaths is clearer.
Change the title: that controls the message
Change the colour to something more neutral.
Pause: notice how the message is completely different.
What can we do to this chart to change the message without even changing the vizzes>
Rotate the viz. Now the drop in deaths is clearer.
Change the title: that controls the message
Change the colour to something more neutral.
Pause: notice how the message is completely different.
Right – here we get into a debate. When you’re designing charts, how do you balance functionality with aesthetic design?>
When designing visualisations for others, you need to balance the ability for the them to actually understand what the heck you’re trying to say with a design that makes them want to look at your chart.
There are so many right and wrong ways to do this it is not possible to tell you what’s right: it depends on your goal. On your audience. On the data. On the questions.
Traditionally we would do visualisation at the end of the process. We’d have our data, and then try to create the chart that demonstrates this. But this was doomed to failure – we’d choose the wrong chart. And we’d stick with it, even if it didn’t necessarily prove our point.
And we’d never visually explore our data to confirm questions, or find new questions.
And who here hasn’t spent days in Excel formatting charts endlessly? We’ve all done it, and while we can pretend we’re doing work, we know we’re really just tweking nothingness.