Ten Rules for Thinking Differently About Numbers

RULE NINE
*Remember Misinformation Can Be Beautiful Too*

I love dazzle camouflage - century-old solution to the problem of camouflaging a ship with a smokestack on an ever-changing sea from u-boats...
The use of dazzling patterns broke up the outlines of the ship, making it hard for periscope operators peering through a tiny 'scope to discern speed and direction. Slow-moving torpedoes would then miss.
Modern dataviz reminds me of dazzle camouflage. Beautiful, yes - but often they misdirect, intentionally or unintentionally.
For example, this CDC graph unwittingly induced an epidemiologist into saying something that wasn't true. The problem is the buckets in the horizontal axis. They're different sizes. Alas, all too easy to make a mistake.
Dataviz is a dual-use technology. It can beautifully clarify and communicate a complex set of data. Or it can obscure, either through bad design or bad motives.
In this recent episode of #CautionaryTalesPod, we hear how Florence Nightingale (played by Helena Bonham Carter) used a souped-up pie chart to start a public health revolution: timharford.com/2021/03/cautio…
And for more on dataviz crimes and triumphs, see "The Data Detective" / "How To Make The World Add Up"
timharford.com/books/datadete…

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Tim Harford

Tim Harford Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @TimHarford

12 Mar
Has the clock of economic history started to run backwards? ft.com/content/c2659b…
Adam Smith famously attributed the "wealth of nations" to the increasing division of labour. People specialised in narrower, more refined tasks, giving us three advantages...
Specialisation boosts output, said Smith, because
1) We perfect our skills;
2) We avoid the distraction of task-switching;
3) We use, and even invent, specialised equipment.

But modern office work fits uneasily into this picture.
Read 10 tweets
11 Mar
Ten Rules for Thinking Differently About Numbers

RULE EIGHT
*Don't Take Statistical Bedrock For Granted*

I begin the chapter of my book on this topic with an extraordinary tale about a steaming-drunk congressman barrelling around the DC Tidal Basin with a famous stripper...
...but really, the story doesn't require that kind of hook. Solid statistics are the bedrock of understanding a complex world, and of holding governments to account.
This recent piece - comparing statistics to radar - gives a sense of what is at stake: timharford.com/2021/03/why-in…
But you can also tell something about the value of independent statistics by looking at the sinister efforts by powerful politicians to intimidate, smear or simply murder independent-minded statisticians.
Read 4 tweets
10 Mar
Ten Rules for Thinking Different About Numbers

RULE SEVEN
*Demand Transparency When The Computer Says 'No'*

Some people will tell you that the world is run by all-knowing algorithms.
Other people will tell you the world is run by crappy biased algorithms.

The truth is...
...that some of them are very good indeed, and some of them are crappy and biased.

What's alarming is that we have no way to tell. Few algorithms have been rigorously proven in transparent, replicable randomised trials. Many are commercially confidential.
Many of the algorithms that seem to be flawed or biased have only been exposed after a tremendous amount of spadework by investigative journalists. I mean, just look at THIS from @propublica propublica.org/article/how-we…
Read 7 tweets
1 Mar
Ten Rules for Thinking Differently about Numbers...

RULE ZERO
*Indiscriminate doubt is at least as dangerous as indiscriminate belief.*

People talk about the weird things conspiracy theorists believe, but think about all the things they first have to disbelieve...
This has been well understood since the 1950s (Big Tobacco starts the 'manufacture doubt' strategy), through climate change denial, negative campaigning - or consider Steve Bannon's famous remark: "flood the zone with sh*t".
Doubt is powerfully corrosive.
It's why I am guarded about Darrell Huff's book "How To Lie With Statistics": it feeds the narrative that all statistics & all evidence are inherently dubious and can therefore be dismissed if we wish.
"The Data Detective" is intended to be an antidote.
timharford.com/books/datadete…
Read 4 tweets
5 Feb
Storytime!

In 1935, officials in the British Air Ministry were worried about falling behind Nazi Germany in the technological arms race...
1/
They decided to explore the possibility of a DEATH RAY.
They had been offering a £1,000 prize for anyone who could zap a sheep at a hundred paces. So far, nobody had claimed it.

2/
Unofficially, they sounded out Robert Watson Watt, of the Radio Research Station.
And he posed an abstract maths question to his colleague Skip Wilkins.
3/
Read 19 tweets
4 Feb
A loyal audiobook listener pointed out that, brilliant, clear and strangely moving as my performance is, there are no pictures in the audiobook of "How To Make The World Add Up" / "The Data Detective". True.
So some links and images below...
1/
Always loved this picture of dazzle camouflage. Modern infographics are a bit like dazzle: they attract attention, then confuse or misdirect...
en.wikipedia.org/wiki/Dazzle_ca…
2/
This infamous image by Nigel Holmes from Time Magazine is a splendid example of what Tufte dismisses as "ChartJunk". Like it? Hate it?
visualoop.com/infographics/d…
3/
Read 9 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!