For this week, Andy once again delights us with a very small dataset: 1 dimension, 2 measures, 7 rows of data. I can already sense the frustration that some people will feel because how big and elaborate a dataviz can one build with only 14 values? I am, on the other hand, delighted. A great dataset to keep things simple. As simple as possible.
So let’s look at this week’s challenge. It’s an article by Elena Holodny in Business Insider UK discussing the 2010 (Q1) to 2016 (Q3) growth of employment in G-7 countries. These include Canada, France, Germany, Italy, Japan, the United Kingdom and the United States.
The data visualisation in the article shows the following pie charts, which depict employment growth as Share of Total Employment and Share of Net Growth in Employment with each country being a wedge of the pie.
What I like about it:
- the colours for each wedge are very distinct, so it is easy to differentiate the countries based on colour
- the labels are large enough to be legible
- the title is pretty clear, identifying what (employment growth) was measured, in which context (G-7 countries) and for which timeframe (2010:Q1-2016:Q3)
- the headings for each of the two pie charts also specifies which measure is being shown and that the numbers are percentages, which by including in the heading saves space on the chart
- I’m glad they included the note on rounding to ensure people understand the ‘limitations’
- the data source is stated
What I don’t like:
- the formatting of the title and individual chart headings looks cluttered. Keeping the font regular instead of bold, or using a lighter font would have helped, as well as leaving more white space – it’s all a bit too dense for me
- it’s a pie chart, with 7 wedges. It’s not the easiest for viewers to visually comprehend the proportions attributed to each G-7 country
- the labels for the small wedges sit off the side of the chart, which looks inconsistent
- having the pie charts next to each other almost feels like we should be comparing them like a ‘before and after’, but they need to be looked at separately, because they visualise completely different measures that don’t have a direct relation
- the branding/logos take up valuable space. Are they really necessary? Would a small icon in the corner suffice?
What I did:
- after playing with the data and reading a bunch of articles to get a better understanding of it, I decided to settle on a simple bar chart to just show the differences in percentages between the countries
- I wanted to finally design a mobile viz, so I went fully mobile rather than adding different device layouts
- I’ve also grown quite fond of large labels and dual labels on a single mark, so I spend a bit of time trying out different approaches until I settled on something I liked
- and because I created a new colour palette this week, I chose to use a different colour for each country. In part that’s also because I divided the charts across 2 mobile screens (when you scroll down), so I wanted to make it easy for the viewer to recognise the countries based on colour.
(Click for the interactive version)