100yrs of Data Visualisation best practice

All updates will be done here:

If you foster a data-driven culture from the top down, you will reap the rewards!

Predicting the need for analytics. Although I hope execs aren’t still being killed off by the stain of their jobs!

Brinton’s world was a man’s world.

Here, he’s not writing about boxing. He’s discussing the menial jobs done by “girls”. 

In the book, the word “man” or “men” appears well over 100 times. 

Woman/women? Zero.

Girl/Girls? EIGHT times.

Check out this page on executives to show just how strong this opinion was (go read it here):

Brinton was ahead of his time. In designing the chart above, he wanted to acknowledge the engaging power of illustration but preserve accuracy. 

If you’ve seen charts like this before, it’s because they are almost identical to the Isotype “pioneered” in the 1930s by Otto Neurath.

Brinton’s motivation was to avoid these kinds of pitfalls:

The charts existed in 1914, but the name’s hadn’t been applied. Here’s a Shotgun chart (p201). Wouldn’t it have been more fun if we’d stuck with that instead of scatterplot?

In 1914, the horror of pie charts was known, even if they hadn’t been named yet:

While Brinton was passionate about getting things represented accurately, he wasn’t fundamentally against a cartoon approach, such as the one above. He said 

“the cartoonist style should not be broadly condemned, for it has tremendous possibilities. There is a great opportunity waiting for the man who can combine cartoon methods with accuracy of numerical statement.”

Getting this balance right is one of the bigger challenges as we come to the end of the Infographic era, and it’s still being argued.

How had analytics developed in the years up to 1914?

The railroad companies in the US were some of the first to tabulate data in annual reports

Prior to this, the figures tended to appear in the paragraphs themselves. This was deemed impractical as it was hard to interpret the numbers:

By the end of the 19th century, people were progressing to charting. The chart below is claimed to be the first chart used in managerial reports. It’s from the mid 1880s:

How do you compare different periods or different categories in 2015? Simple - you move your data around and explore it. The GIF below shows various ways of seeing seasonality.

Back in 1914, Brinton recognised this too. But how could you implement it? 

Index Cards! They can go side-by-side or one atop the other:

And don’t worry, Brinton has a whole chapter on how to implement, store, maintain and secure this system. Fret not, “it is work any man of even ordinary caliber can do”

Isn’t this what we’re STILL trying to get right in 2015? What’s not working?

It was this frustration that inspired Brinton to write his book in the first place. Here’s his opening comments:

How do you know your viz is complete? Well, you could check this 100 year old checklist. Brinton’s list pretty much remains valid today

(go read the whole list here)

There are some fascinating technology issues that have changed, though. For example, we don’t need to consider Ben Day dots any more, fortunately.

Don’t size by area, said Brinton. He really bangs on about that. In general, he’s right, because we cannot compare area accurately. However, is it fair that sometimes area is more engaging than length? And if a sense of the data is all that’s needed, is that ok? Here’s one from the Economist where area is engaging.

But sometimes it just doesn’t work:

Brinton cautioned against doing crazy designs because the eye is so easily fuelled. He had several examples of optical illusions, but I like the one with the old-fashioned copper and children.

There is an excellent post on Visual.ly about this you should read to get the modern perspective: