New ways to visualise time

[Updated for Tableau Conference Europe 2018]

I’m delighted to present “New Ways to Visualize Time at TC Europe in London. The content is based on one of the chapters in my book, The Big Book of Dashboards.

This post contains links to all the resources shared in the session.

The trend line is amazing. It shows peaks and troughs and trends. But if you only ever use trendlines to show time, you are missing insights in your data. What’s the best way to show time in visualisation? I cannot answer that: it depends on your data, the story you want to tell, your audience, and many other things.

CollageWhich way is time?

Time can go up and down as well as left and right. Generally in the West we point forwards or to the right, but that’s not the only way. Check out these:

  1. For further reading on how time is considered in culture and in data, check out this wonderful article, History on the Line, by Stephen Boyd Davies.
  2. Some dataviz goes in unusual directions:
    1. The Nasdaq viewer by WSJ goes forwards
    2. XKCD Global Temperature goes downwards
    3. Mike Cisneros’ Oil v Gold is an excellent Tableau example of vertical time

 

Jacques Barbeu Dubourg's Chronographie from 1753
Jacques Barbeu Dubourg’s Chronographie from 1753

A history of the trend line

We forget (or don’t know) that even the most common chart types were once ideas waiting to be thought of. Even though we build them every day, timelines were invented. Here are the milestones I highlight in my talk

For more reading, I cannot recommend Cartographies of Time highly enough. It is an amazing book about the history of the timeline.

I also love this fascinating article about time as a dimension in which we find the quote “the mapping of time has made only modest intellectual progress since it was invented 250 years ago.”

6 new ways to visualize time

Slope chart

Cycle plots

A cycle plot (or is Batman?)
A cycle plot (is Batman hiding in the middle?)

Highlight tables and small multiples

Index chart

Bump Chart

A bump chartA bump chart

Other resources

  • One of the examples I used is based on US Road Fatalities data. I used that data to create a dashboard that was comeprehensively described and deconstructed in my “Design Month” series of posts in 2014.
  • I wrote about a similar topic for the Huffington Post, “New ways to see time.” It has some other examples.

 

 

Lollipop charts, revisited

Easy on the eye

Stephen Few has written about lollipop charts in his latest blog. “Malformed bar graphs,” he calls them. The poor things – lollipops are more than that.

The problem is, he didn’t consider the problem I was trying to solve when I first created them in 2011*:

Problem: A bar chart with many bars of a long length are unpleasant to look at.

My eyes! They hurt!

I believe lollipops create a visual experience which is easier to look at. In his post, Steve used an example using 7 bars with a large range. I didn’t create the lollipop for that situation. While I believe a bar or lollipop works for that kind of data, criticizing the lollipop without addressing the original intention is disingenuous. I did consider fat bars and thin lines, too, in seeking a solution. Fat bars are just too inky, and the thin lines don’t have enough definition. Lollipops are an attractive compromise to solve the problem.

*Did I invent lollipop charts? Alberto Cairo credited me with their invention in his book The Truthful Art, and I’m not going to argue against that! The images on my original post are no longer available, unfortunately.

 

Changing the message without changing the data

Two formats, two messages. Time for a new example?

If you’ve seen me present in the last 3 years, you’ll probably have seen me show the Iraq Bloody Toll chart. Then you’ve seen me turn it upside down to create an entirely different message (full post here).

I still love showing this example to new audiences. I love seeing the light bulb go off as they realise that a data and a chart is just a method of communicating a message: facts are not neutral.

But it’s time to find a new example and for that I turn to you for help.

Have you got any other great examples of charts where the message can be transformed in as simple a way as this one? 

(Note: I’m only looking for examples that stay true to good practice. Truncating the y-axis doesn’t count!)

There are some older examples. Obama’s bikini chart was a cracker, described very well by Robert Kosara in 2012.

Do you know any others? If I can find enough, we could turn this into an entire blog post or webinar. Let me know in the comments or on Twitter.

The little of visualization design: grouping with colour

[I’m shamelessly stealing the idea for “the little of” from Andy Kirk. His series has been amazing]

Yesterday in MakeoverMonday we tackled the worldwide shipping industry. Today the Economist shared a chart about the same topic (above). I highly recommend you also read the full article as it’s about an industry trying to save itself through the use of better data.

What’s I like in this chart is the “Proposed alliances” blocks next to the company names. The primary question in the story is the size of the shipping companies. The secondary question, dealt with in the second half of the article, covers alliances. They designed the chart so you can get on with answering the primary question without the initial distraction of the alliance information.

If I’d designed this chart, I’d have probably coloured the whole bar according to the proposed alliance. I’ve mocked that up below.

coloured-bars

What’s the problem? You can’t escape the colours. You see the colours before you answer the primary question. What the Economist did was keep simple bars that allow the primary question to be answered first. Once that’s dealt with, you can examine the alliance colours and focus on the secondary question.

The perils of bubbles

Another day, and another tragic event. The events in Nice horrified us all. One excellent response is to reinforce that, should this turn out to be a terrorist attack, the individual does not represent all Muslims. The number of terrorist-minded Muslims is tiny. We can make that point using data.

This afternoon I saw the following chart tweeted by Ian Bremmer, which makes this point:

The point being made is great, but look at those circles: if you put that ISIS circle inside the Muslim circle, it’s kinda big. So I took the data from that chart, and redrew it in Tableau. Turns out, the circles were the wrong size. Here’s what the size really look like, if you draw the circles so that their AREA represents the value:

Muslim

 

Can you see the circles for Al Qaeda, ISIS, or the Taliban? They’re up there in the top right. Tiny, aren’t they?

The lesson here is that, if you’re going to use circles, size them according to the area. The people behind the original chart, MIIM Design, were trying to make a valid point, but the circle size misrepresented the sizes, which could cause confusion.

Note: I used the same numbers MIIM used in their chart, taking the upper estimate of each category. I have not done research to check the validity of these numbers. My goal is to make a point about circle sizes, not a political point about the size of different religious or terrorist groups.

Update: I replaced the original image with one with a new title. One commenter suggested, rightly, that I might have implied “Muslim” is a terrorist organisation. That absolutely was not my intention or my belief.

 

Dataviz criticism: know the author’s intentions first

Which do you prefer?
Which is most effective? Which do you prefer? Cole on the left, Steve on the right?

Stephen Few has been discussing 100% stacked bar charts. He’s opened a great debate with Cole Nussbaumer. I recommend you go and read the blog and the comments. In a nutshell, Steve prefers lines over stacked bars because they allow comparison of all data points at all places in the time series.

My problem is Steve’s assumption that Cole intended people to be able to compare all data points at all times in chart. In one of the comments, he says this:

“The information and eventual understanding that we get from a graph is more important than any impression that it provides unless that impression is all that matters.”

What if ‘that impression’ IS all that matters?

What impression do you get from this?
What impression do you get from this?

Sure, Steve’s lines allow you to do more accurate comparison of more data points. However, why should that be the only purpose of a chart? That purpose is one he imposed on Cole’s chart. It might not be the objective Cole had in mind when designing it.

Maybe Cole’s objective was to highlight the % missed only (ie the red marks on each chart). The other bar segments are secondary information, not vital to her key objective in making a 100% stacked bar. The “impression” I get from her stacked bar is that Missed Targets have increased to 42%. If her intention was that I get that impression, then it’s a success.

I might argue that if that was her intention, she could have drawn just the % missed data, and ignored the rest. However, this is at the expense of secondary information. Here’s what that could look like:

Show just the data necessary
Show just the data necessary

Steve also wanted other example of multi-segment stacked bars. I found one from my UK Election project:

% of total

I acknowledge the problems with this chart: it’s very very hard to see which organisation tweeted most about UKIP, or the Lib Dems, or any of the other central segments.

But my intention was NOT to allow comparison of every segment. My intention (as shown by the title of the view) was to highlight which organisations were tweeting most about the Conservative party. That was my prime goal. The Conservatives are the left-most segment and sorted in descending order: it’s easy to see which orgs tweeted most about the Conservatives.

Other information which can be learnt from this is secondary to my prime purpose and therefore was intentionally compromised by using stacked bars. An interactive version would add tooltips and more contextual data, allowing the curious to discover more in the chart.

I acknowledge the stacked bar isn’t perfect, but I don’t know how else I could have designed the chart so that it answered my prime intention (% of tweets about Conservatives) and allowed the viewer to see secondary information. How would you have redesigned it? If you wish, the data is here.

Remember, every visualisation is a compromise. And every visualisation has a prime intention which must be considered before critiquing it.

I wholly recommend the following posts for further reading:

 

Design Tips for Functional and Beautiful Dashboards

before after portrait

Over on the Tableau blog, we published a post on how easy it is to quickly format a key metrics dashboard. We then redesigned it in response to some valid criticisms of the original (click here). The changes made are subtle but effective. In this post I wanted to describe them to show how to maintain functionality when going minimal with their dashboards.

Example 1: the line charts

before and after line charts

What was wrong with the original? The lack of labels make things clean, but stop you learning anything about the values:

why label

I changed the following:

  • Added the x-axis headers. I’m all for hiding headers (see “How to design an axis for maximum impact”), but I think it’s wrong to hide the x-axis.
  • Dual axes provide better design control in Tableau:
    • The Sales chart is an area/line chart. This gives more definition to the top of the area.
    • The profit chart is a line/circle chart. I find this gives more definition to the points than just choosing the “All markers” on the Colour shelf:
      Dual or all Markers
  • I wanted to avoid showing a y-axis but there still needs to be some way of seeing the magnitude of the measure. For this I added a maximum reference line, with the text aligned to the right. If a user filters the dashboard, the max will always be accurate
  • All titles were changed to be left-aligned and single line. This creates better consistency across the dashboard. I aligned the Titles to the left and the Maximums to the right in order to prevent visual clutter and confusion. If they were closer together, it’s harder to tell which is which:
  • Note also that the zero line on the profit chart above is a Constant Reference Line, rather than an actual axis. I did this because it’s easier to control the formatting on a minimal chart.

Example 2: the Profit Charts

donuts

I don’t object to donut charts in all circumstances, but they are simply a bad choice when you make a donut for each Year (see p11 of Stephen Few’s classic, “Save the pies for dessert”). It’s nigh on impossible to see changes over time. I switched to a stacked area chart instead. This took up less space and allowed me to label the marks, too (removing the need for a y-axis).

Notice how the colour for Furniture is a really light salmon pink. That can be a problem against a white background, but I added borders to the marks to make them more obvious:
borders

BTW, I didn’t address a key risk with this kind of chart: what happens if there is negative profit?

Example 3: the highlight table

highlight table

I love highlight tables. However, I didn’t think this one worked for 2 reasons. First, the lack of labels creates a real problem here. Without them, it’s just a mosaic. I also was not convinced that you’d ever really want to see Sales by Week and Weekday. I changed it to Month and Day. The increase in sales towards the end of each year is now clear.

Example 4: General changes

  • Colour changes
    • I used extensive use of grey in the text in order to soften the labelling and accentuate the data marks.
    • I used more distinct colour palettes to differentiate Sales and Categories.
    • I switched from a floating layout to a grid layout. This way, Tableau controls the size of the grey borders between the charts.
    • I have to say, changing the colour was the single hardest thing on this. All the other decisions were kind of straightforward but every time I changed the colour, I felt like I’d made one thing better but something else worse. Also – the image on the Tableau blog is from before I added the borders around the marks. 
  • Layout
    • I changed the layout so that Sales-related charts are on the left, and Profits ones are on the right.

Conclusion

Would you have done something different? The dashboard is still not perfect. For example, the blue highlight table and the blue map are problematic. However, I hope this exercise has shown that it is possible to have both beauty and functionality in a Tableau dashboard.

The key lesson for me is that it’s not hard in Tableau to acknowledge feedback and then iterate fast to fix the problems.

Feel free to download the workbook and share your ideas.

Before and after...
Before and after…

Carlisle Flood Defences: the one chart I’d like to see

Carlisle flood

Carlisle, in the north of England, has been hit, again, by severe floods. I feel huge sympathy for those whose homes, again, have been damaged by floodwater. The £38 million spent on flood defences by the Environment Agency since the last major floods in 2005 is being described as ineffective. “How could all that money not have protected the houses?” they are being asked.

First of all, the EA never said the defences would protects against all future flooding.

What I’d like to see is a chart like the one above, but with real rain data. All the labelled years were genuine flooding years, but I do not know the actual rain values (apart from the 34.1cm-in-one-hour being reported for 2015). For the purposes of this, I’m using that value of 34 as the index value for 2015.

When setting a flood defence, an agency has to balance cost against the predicted maximum future level of flooding. If the EA predicted in 2005 that there would never be a 34cm deluge, then they wouldn’t waste public money building flood defences that high.

With real data in the above chart, we’d be able to see:

  1. How unprecedented is 2015? (in my chart, it’s 5cm higher than 2005)
  2. What level of rainfall are the flood defences set to protect against? (in my chart it’s higher than 2005 and 1822 flood events)
  3. What predictions do they have for future precipitation levels and how regularly will they be higher than the flood defence level?

If 2015 truly is unprecedented and wholly exceptional, then blame shouldn’t be put on the flood defences.

This post was inspired by memories of the Fukushima disaster. Then, the tsunami which hit the power station was higher than the station’s sea wall.

What do you think? Would this chart help asses the success of the Environment Agencies flood defences? What chart would help you answer the question?

NOTE: don’t forget, the data in the chart is not real. It’s for illustrative purposes only.

Is information visualisation research flawed?

From "Beyond Memorability...."
From “Beyond Memorability….

Stephen Few’s latest newsletter, “Information Visualization Research as Pseudo-Science” is a critique of the academic process in visualisation research. In it, he savages one paper in particular: “Beyond Memorability: Visualization Recognition and Recall.” He uses this as an example of what he thinks is a problem widespread in this field.

I agree there are problems in this paper. I agree with his suggestions for fixes.

However, I think it’s unfair to say this is a problem with visualisation research: it’s a problem with all research. In all fields, there are great studies and there are bad studies.

In this post, I’ll explain my own thoughts on the flaws of the paper, then the areas where I think Stephen is being unfair.

Here are my own thoughts on reading the paper (which I noted before I read Stephen’s article, as he instructed):

1. Stop publishing 2-column academic papers online!

Not the way to read papers online!
Not the way to read papers online!

Why are academic papers STILL written in two columns? This is ridiculous in a time when most consumption is on screen. To read a 2-column PDF on my phone or tablet I need to do ridiculous down-up-right-down-left scrolling to follow the text. Come on academia: design for mobile!

2. Why are they measuring memorability?

I agreed with Stephen on the key problem: why are they measuring memorability? Isn’t it more important to understand the message of a visualisation?

3. Hang on, Steve! Problems with experimental technique are not unique to visualisation research

Stephen goes to town dismantling the study’s approach. For example, he criticises the small sample size and much of its methodology. I am not as expert as Stephen in this, but I find myself agreeing with most of this.

But where I differ is how he damns visualisation research as if the rest of research doesn’t have the same problems.

Let’s look at some:

i. Statistical unreliability

There are no shortage of academics papers with statistical problems caused by small samples. Here’s one on fish oil, dismantled by Ben Goldacre. Incidentally, the study he refers to also used 33 subjects.

He also outlines a statistical anomaly so extreme, that half of all neuroscience studies are statistically wrong.

Conclusion? Statistical problems are not unique to visualisation research.

ii. Methodological misdirection

How many of the 53 landmark studies in cancer had results that could be replicated? 6.

Yes, 89% of landmark cancer studies have results which cannot be replicated. (source: this great article “When Science Goes Wrong” from The Economist)

Conclusion? Methodological problems exist in all science.

iii. Logical fallacies

Logical fallacies are hardly unique to visualisation research. For example, this list of the top 20 logical fallacies is a good example of how this is a problem in all science, not just visualisation research.

Part of this critique is surely just part of scientific rigor?

For a conclusion, I acknowledge that I’m not an academic and I don’t read many academic papers, so I am naive.

Part of me thinks that surely lots of this critique is just part of scientific research? Researchers publish papers and the world responds, positively and negatively. Future research then improves.

I assume Stephen’s frustration stems from the fact that many of these problems are perpetual and should have been fixed before the study started. I can’t disagree with that. But I don’t think the paper is “fundamentally flawed” as Stephen describes. Maybe memorability of the view is important? If so, this is a first step in the iterative, slow advance of academic research. The paper at the very least makes us consider the question of what it’s important to remember from looking at a visualisation. Having read it critically, I have considered the question and formed an opinion. That’s of value, surely?

I found it very interesting to sit and really read an academic paper in detail. I don’t do it often, and I respect people who can wade through the dense formulaic wording to get to the meaning.

[Updated 5pm 3 Dec to expand my summary]

 

What’s YOUR elevator pitch for data visualisation?

If you had to pitch dataviz, what would you say?

In this week’s #AskAndy Anything About Data, Chris Love (@chrisluv) asked myself and Andy Kirk the following:

A great question to ask, but tougher to answer than you’d think. Andy K and I both gave our answers (go check the webinar recording to hear them). I then threw it back to Chris. He gamely took on the challenge. Here it is:  

Nice work, Chris.

NOW IT’S OVER TO YOU: what is YOUR elevator pitch for data visualization. Give yourself 30 seconds to 1 minute. How do you pitch data visualization?  Tweet your efforts to me at @acotgreave. Hashtag: #datavizElevatorPitch

2853282
Image from LinkedIn Pulse