When I first started this blog, the Marimekko was swept up in my overview of resources about this chart type. I didn’t really understand the purpose of the chart back when I wrote the post back in July 2016, but that was when I was frantically scrabbling around trying to compile a central repository of posts to refer to when trying to get my head round the product.
The perk of blogging about everything is that I have a reference point to go back and read, which was the entire point of the blog, and pleasingly it seems some people are using it for that purpose too.
Even back when I wrote the Marimekko post, it seemed that Jonathan Drummey‘s excellent series of posts were regarded as “the” resource to refer to for Marimekko’s, and as Emma Whyte referred to them in her #WorkoutWednesday post, it seemed sensible to defer to Jonathan’s guide to have a stab at the challenge this week.
First of all, we need to isolate the “Total” within “Sub-Type” (which I renamed Gender):
Which plotted in a table looks like this:
Those percentages don’t look very percentage-y, do they? And as they’re whole numbers, switching the Number Format to Percentage isn’t going to do us any favours:
The solution is instead to just use a “%” suffix:
Diligently following Jonathan’s post, I openly confess that I have not got a bloody clue what the next bit does until reading his detailed explanation! Not a clue, but his post is so well-written that it’s simple enough to adapt to this dataset:
Here’s the underlying logic, and again it’s articulated superbly – adapted by me for this challenge:
This formula…..is a bit complicated. Here’s how it works:
a. If it’s the first address (aka row in the partition) then return the [Total Percentage].
b. If there’s a change in the [Job Type] from the prior address (i.e. we’re in a new column), then return the sum of the prior value of this measure (from the prior address) plus the current [Total Percentage]. This increments the running sum for each new [Job Type].
c. If there’s no change in the [Job Type], then return the prior value of this measure (from the prior address). This carries forward the running sum across the [Job Type] without incrementing the running sum so all [Job Type]s in the same column will have the same value.
Continuing with Jonathan’s guide, we arrive here:
So everything is working as expected, and now it’s time to move away from validation tables, to charty stuff.
So [Complicated Bit] computes using Job Type. What do we need to do? Figure out how to size the bars, bring the Gender split into the view, and sort stuff so the “progression through the ranks” of Job Type makes clear the key point of this dataset: the fact that the percentage of women declines sharply at higher job levels.
According to Jonathan’s guide, I need to plonk my [Total Percentage] field on Size, and critically you switch from Manual to Fixed Right Aligned Size:
Now I want to bring Gender into the view, so we can carve up those big blue bars and start to make sense of stuff.
Nuts. I get the Total part of the Gender field. But if I exclude that…..
Well that’s a bit wrong. So I duplicated the Gender Dimension and dumped it on Detail as an Attribute, excluding the Total member:
OK. So we have sorting to resolve. So I want to sort by Job Type and Gender. Job Type is to be arranged in ascending order by pecking order, and Gender so the Female Dimension member is lower in the view.
Job Type was a Manual Sort in my case:
And Gender was just a descending sort by Percentage:
With the colour picker and some clean up, I could get pretty close, pre-labels:
All the labelling I just floated as text objects, with the exception of the Image of the arrow, which again was Floated.
The final dashboard is here.