For week 30 we challenged the community to improve on a bar chart published in an OECD report on parental leave systems in countries around the world. We saw a lot of great makeovers for this data set and overall a large number of submissions for this topic. I particularly noticed that we had a number of new female contributors who seemed encouraged to tackle a subject which they cared about and wanted to visualize.
We saw a lot of scatterplots which was not surprising given the multiple related measures in the dataset. I’m personally a big fan of scatterplots and I enjoyed the various submissions that came through.
A great example came from Klaus Schulte who added curved reference lines with shading to his scatterplot to indicate which countries fall into which range, depending on their combination of duration of maternity leave and pay rate for this period.
I also liked seeing many submissions with a single chart. There is no obligation to create a dashboard each time and doing a simple visualization that is formatted well and tells a strong story can be much more impactful than a dashboard with multiple elements that isn’t really saying much.
The master of simplicity this week, and I say this as an absolute compliment, is Charlie Hutcheson who created a simple dot plot which highlights the difference between the USA and all other OECD nations in the list.
LESSON 1: VERIFY YOUR DATA
A common mistake this week was to include all countries in the data set and call them OECD countries. If you look closely at the data in the original article (which we encourage people to ALWAYS read as part of their makeover, that’s why we provide it in the first place), you will see that most of the countries are in the OECD and some are additional European countries included by the authors.
Admittedly, calling the challenge ‘parental leave in the OECD’ may have led people to assume that all those countries are in the OECD, but in your analysis process you should check these assumptions. Had I call it ‘parental leave in Europe’ and included countries like Japan and the USA, would you have questioned the title’s accuracy?
I don’t expect people to know which countries are in the OECD and which are not, but a quick check of the original source will give you the answer.
Then what? Simple, actually: Either you limit your data set to the OECD and use that description or you include all countries and don’t call it OECD.
LESSON 2: DASHBOARD DESIGN
A lot of our Viz Review comments relate to dashboard layout and design so I wanted to share some examples from our webinar and the recommendations we made, regarding layout and colors (we saw A LOT of color this week…).
Each example features the original submission on the left and the revised version on the right.
- make the scatterplot square
- simplify axis labels
- ensure OECD relates to correct list of countries
- add more white space in the dashboard between elements
- bring the BANs to the left to move them more into the focus
- reduce the prominence of the reference lines by making them dashed lines
Example 3: Colors | Naomi J Ball
- remove background color
- connect milestones more clearly to data
- move milestone axis to the left (as we read from left to right and need this information at the beginning of the chart not at the end of it)
Well done to everyone who participated this week. A seemingly simple data set which still required thorough analysis as well as clear communication of the findings to ensure the audience was presented with a clear picture of maternity leave systems in different countries.
Here are this week’s