Iron Viz Europe: Changing Demographics Across Europe

6 days ago, I tweeted out this picture of my work-in-progress dashboard. Below, is my narrative on how it was built.

Iron Viz season is well, well, well underway. For the first time, Tableau Public are running Iron Viz Europe - Think #TheBigOne, but on a smaller, more concentrated scale.

I entered my first ever Iron Viz for the first feeder this year, looking at National Parks in the UK, but the topic for the inaugural Iron Viz Europe was simple; Europe. Any data, any topic - but it has to have a link to Europe.

Now, there's actually a surprising number of data portals in Europe, and quite a few with good, clean data. I submitted a piece of work I did a few months back on EU GDP Growth, and following a similar thread, I decided to focus on something based around the EU28 countries.

I'm a semi-regular reader of the Economist, and one article which I read last weekend caught my eye - "Immigration cannot plug the hole in America’s budget". Within this, it detailed that America's ageing population would mean greater pressure on the working age folk to take on responsibility + productivity - and to solve this with immigration, I quote..

A rough calculation suggests that almost 40m such immigrants would need to arrive immediately in order to fill the hole caused by social-security payments and hospital visits for the over-65s. They would have to be followed by 36m more by 2047—arrivals that are already baked into budgetary forecasts.

Now this lead to a discussion with my uncle, who quoted something about how fertility rates would have to be maintained around 2.0 to sustain a culture through a generation. Coupled with the article, I wanted to take a look deeper for myself.

For this, I used the World Bank Dataportal. This has a whole host of development indicators and macroeconomic data open for use. I took a handful of metrics, and then got to work. Through the #TableauShuffle, and some data exploration in Tableau, I figured what I wanted to show is how much things had changed. The data I had extracted I'd taken from 1984 (as this gave the most complete data for all the EU28 countries) up to 2014/2015 (some metrics didn't have data for 2015). Small multiple with text was my original idea, and I very quickly had a rough visual in my head of what I wanted to do.

A calculations I use often, but can never remember.

First, I thought of using the "Columns/Rows" calculations (which I save in my OneNote cos I can never memorise it) [For reference; Column Divider: (index()-1)%(round(sqrt(size()))) Row Divider: int( (index()-1)/(round(sqrt(size())))) - then make both discrete dimensions and have them compute using the dimension you want to split into small multiples).

Hmm.. Not enough whitespace for me

This trick is one I love, but it was too crowded for me. I wanted my small multiples to have space to breathe. So I took 2 unprecedented steps;

1. I decided to use floating containers.
2. I decided to create a long form dashboard.

I've been (perhaps fairly publicly?) against long-form visualisations, mainly out of laziness and effort placed on the reader. I think it's a difficult skill to build out a long-form viz, while also maintaining a strong rhythm - particularly balancing an emphasis on design vs and emphasis on information.

Nonetheless; I wanted to do small-multiples here, and I wanted to tell a story. So I began designing my dashboard and working out exactly how long my dashboard should be, how many elements I want and how to lay it out. The tools I used for this were; Microsoft Excel to plot where 28 countries would sit, and how much space a title, a line chart and stacked bar would need.

How to scientifically figure out what goes where on a floating container dashboard

Magical.

Now here's the interesting part. What I started with is an idea where I had the EU28 average in the middle, and all the countries around it. But this was waaaay too fiddly, and I wouldn't be able to use one y-axis as I can with this finished product. Below, you can see what I started off with as my first pass.

As things moved on a little, I found that I needed a touch more space in my visualisation. So, given I've plotted the exact position of my containers, said exactly the height, width and layout - These should all be fine right?

Wrong. Tableau shifts all these down - I've added it to the list of questions I have when I grab a dev at the Tableau Conference in Vegas this year. That, and a font issue I'll cover shortly.

As you can see, layout, colour scheme and background have all changed - and this is thanks to the waves of feedback I got; not only from those in the Data School, the Information Lab (UK and beyond!) but also from members of the Tableau Community. Y'all rock. Thank you!

The final touches for me was to clean up the "how to read" legend - Ben Moss was a big help here by suggesting I combine the image I'd mocked up with the EU average visualisation, thus leaving me with more space to play - and the final three panels at the bottom. I called these "quick hits" - thought they're more like little blogshots. As as well as changing proportions within the EU28, I wanted to focus upon the fertility rate (ie. how many babies we're having) and the dependency ratio - ie. which countries' labour force have the highest burden upon them.

Finally; the font. I used Segoe UI throughout this, however I noticed some strange behaviour when I uploaded to Tableau Public - The text boxes, titles, are all Segoe UI, but the axes and labels seem to revert to Times New Roman (or whatever the default font is) - This lead me to change it to Arial, for consistency.

Conclusions

This is the first time I've used floating components in a dashboard and created a long form viz. While I understand the reason that they are popular, as a designer I don't have the patience or the mindset to regularly create these - particularly given the control (or lack of) that Tableau provides.

While I can't envision that I'll create a long-form for a long, long while - I'm glad I pushed myself to try something different for this Iron Viz entry; particularly for a topic and story which I am fairly passionate about.

To view the viz, follow this link; http://tabsoft.co/2p58rM0

Or check out the image below.