Finding a Story in Europe’s Over-qualification Rates

What is the data about?

The data used in this post relates to information brought together by Eurostat in their Experimental statistics on ‘Skills’. Specifically, the “over-qualification rate by economic activity for the period 2008 to 2016”.

Rather than paraphrase the Eurostat definition, you can read it here. However, to briefly set out why we thought it was an interesting dataset for #VisualisingHE to have a quick play with, we’ve taken a few sentences from the Eurostat preamble about why we may need indicators about skills:

The European Commission communication ‘A New Skills Agenda for Europe‘ defines among others policy priorities and sets out actions to be undertaken with the aim to make better use of people’s existing skills. It should be ensured that the skills available in the labour market match the needs of businesses and the economy.

This undertaking includes finding ways to improve the matching between skills and labour market needs as well as bridging the gap between education and work. Skills mismatch indicators should measure the gap between demand and supply of skills (macro-level) as well as conditions of workers, jobs or vacancies (micro-level).

Where does the data come from?

The indicators are derived by combining available figures using EU Labour Force Survey (EU-LFS).

The use of a single source ensures consistency, comparability and reliability, and the methodology behind is based on the literature in the field (‘The skill matching challenge: Analysing skill mismatch and policy implications’ – Cedefop 2010) and on existing empirical exercises (‘Employment and labour demand’ – Eurostat 2016 and EU-LFS ad hoc module 2000).

Full methodological description including data.

What does the original data look like?

The image below is an example of one tab of the 8 occupational job categories, 2008-2016 by European country: Wholesale and retail trade, repair of motor vehicles and motorcycles


How might we present this data for better consumption?

Having brought the data together using Easymorph, Adam wanted to understand the distribution in proportions of over-qualified roles by occupation types for a single year (2016) to establish which sectors where hotbeds for skills mismatches across Europe and whether the data presented any areas with differing deviations.

The following chart highlights Transportation and storage as the occupation type with the highest median proportion surveyed, with Spain the country with the highest rate @ over 70%.


[Link to interactive viz]

This visualisation of the data provides the ability to grasp all of the proportions for all of the countries, for all of the occupation groups in one visual. The interactive viz also enables you to delve into where countries appear in the distribution of proportions.

How might we present the idea of change over time?

Wanting to recreate elements of the original data, a focus on the percentage point change in proportion over qualified between 2008 and 2016 followed:


[Link to interactive viz]

Again, presenting the change for all countries and all occupations using the box plots enables you to contextualise a particular countries change against that of the other countries.

Focus on a single occupational grouping

Could we create a simple highlight table only, the success of this is wholly down to Chris Love of the Information Lab, he helped me work through the calculations needed to rank by a single year, which turned out to be a little more complicated than initially thought. blog post here.

highlight table

[Link to interactive viz]

How could we present all the data in one visual?

Could we go one further and present all of the year data, all of the country data and all of the occupational data in one chart? And what would you get out of it?

Having a bit of fun with presenting change with Radials?  This view does draw out the changes so that you can see that wholesale and retail have had an increase in virtually all countries where as something like construction has had some big decreases and a fair few more moderate increases.

radial of change

[Link to interactive viz]

What about trend over time?

Elena took a more focused approach and only produced one visualisation that expands on the idea of change over time.

By utilising a slope graph and Tableau’s latest out-of-the-box style if step line charts, Elena’s dashboard allows the user to see how a country’s overqualification rate changed from 2008 to 2016 in each of the eight occupational areas whilst also allowing the reader to see the fluctuation over time. The colour indicates whether the rate for the country in the given year was above or below the median.

Overqualified in Europe.png

[Link to interactive viz]

So….. Did we find a story?

The data threw up plenty of small nuggets but no overall narrative so for the moment this post will be a demonstration of what data is there and how you can visualise the data in a few ways to present the proportions and change in proportions in more ‘graspable’ ways.  This also allows the viewer to identify some of the outliers and find trends in either a particular occupation type, or country.

Thanks for reading!

Team #VisualisingHE

(doing a post on something a little wider this month – Education, moreover over education!)


Destination: Europe

In today’s globally connected world we experience the notion of a border mainly when we go through passport control at the beginning or the end of a journey. For the education sector, it means that now studying abroad has never been easier, especially for those who live in Europe. So what does the student mobility within Europe look like?

To answer this question, I have decided to look at the Higher Education data from Eurostat and find out the number of enrolled students within each European country who are from a different European country. The full dataset contains student enrolments from some countries outside of Europe, too, but for the purpose of this exercise only European origin and destination countries have been included.

Why visualising student mobility within Europe?

Choosing the topic wasn’t easy. In fact, Dave Kirk helped me with this a few months ago when Dave, Adam , Stephen and I launched #VisualisingHE. He and Adam were enthusiastically listing topics and datasets they were looking forward to exploring as part of the project but I was struggling to come up with something that I could talk about with so much excitement.

Big Map 3.png

Then Dave asked me: what are you passionate about? And it didn’t take me long to come up with an answer – I am an Eastern European and I cannot imagine who and where I would be today if I hadn’t been given the opportunity to undertake my bachelor’s degree in the UK! Therefore, I decided to look into the topic and  develop a pretty, colourful and informal visualisation that represents the beauty of studying abroad… in Europe.

What influenced the design and choice of charts?

Probably the better question here is not ‘what’ but ‘who’ influenced the choices I have made as part of the design process. Having decided on the overall ‘feeling’ I wanted to my visualisation to convey, I had to think about whether I wanted my dashboard to be in a #coffeetableviz (i.e. poster) style or be leaning towards the exploratory side. I will leave it to you to decide where precisely on the informative vs. exploratory spectrum the end result lies, but I will let you know whose work gave me the inspiration:

  • for poster style infographic I would always go to Pooja Gandhi’s fantastic work first: if you haven’t visited her tableau public profile, do it! I find her work truly inspiring – she knows how to deliver a clear message in a dataviz by seamlessly combining text and graphs.
  • for visualising the inflow vs. outflow concept I was inspired by Neil Richard‘s viz on the Eurovision song contest. When I first saw his viz, I remember thinking: ‘Oh, well it is clear to me what the visualisation is telling me.’ Plus, I really liked the flags.
  • for colours I got inspired by Jeffrey Shaffers Beautiful Trash.

Then the idea of a Sankey sprung to my mind and that was it – I was going to do a Sankey and practice my great arc curves on a map. I started with the great arc curves as I had already done these for a project at work. For the data structure I followed this summary and to get the lines curved, I found this post useful. Building the Sankey wasn’t too difficult either – there are plenty of blog posts written by the community that explain how to do it. I read a few and after deciding I did not want to use custom SQL, I found this post helpful.

What were the challenges?

Just as I was thinking I was making some really good progress, I came across Ravi Mistry and Nicco Cirones very interesting work on ERASMUS. It was great! I loved it… but it also had a map of Europe with curves, and a Sankey, and was talking about students studying in Europe… Oh dear, I didn’t want to copy their work… Even though I hadn’t seen it until I was alredy halfway there with my viz, if I were to use exactly the same combination of graphs would have still made me feel as if I had copied it. At this point I had two options: drop the topic, or think really hard how to make my visualisation different. A few days later I was going through some of my favourite dashboards on Tableau Public and then, inspired by Adam McCann’s beautiful Beatles Analysis, I decided to develop an ‘hourglass’, or a double Sankey, chart (I don’t know if this chart has a name) to show the inflow vs. outflow relationship. Not long after I had my first draft:

Student Mobility

Another big decision making point in the process of developing this dashboard was around my choice of font. It crossed my mind that there is this thing about safe fonts for tableau public. After a quick check on Google I thought I should try and use a web safe font. My choice was Comic Sans. Yes, Comic Sans, because I was going for a fun artistic / informal look. Little did I know that there was something about Comic Sans… When I asked for some feedback on the way my dashboard was coming together, I got mixed reactions. I particularly like the contrast in Dave and Stephen’s comments:

Comic Sans & Fonts

After doing some reading on the internet and quite a few lengthy discussions with friends and colleagues at work, I decided that though I see nothing wrong in using Comic Sans for my dashboard (especially, since I am not saying: ‘Danger, Danger, the world is about to collapse’ … or something in those lines), I decided to change it. Simply because I didn’t want the font to distract from its overall purpose.

What did I learn?

  1. How to build a Sankey
  2. How to build an hour glass visualisation
  3. How to crop an image in GIMP (and add transparency to it)
  4. There is something about Comic Sans…
  5. How to use A LOT of containers (with very few of them floating!)

What are my personal challenges for the next viz?

  • Don’t use a long-form dashboard
  • Produce a dashboard with a non-black background

Thank you for reading and I hope you enjoy exploring the dashboard!


Inflow & Outflow (6)