A twitter user sent me this chart from Germany.
It came with a translation:
"Explanation: The chart says how many car drivers plan to purchase a new state-sponsored ticket for public transport. And of those who do, how many plan to use their car less often."
Because visual language should be universal, we shouldn't be deterred by not knowing German.
The structure of the data can be readily understood: we expect three values that add up to 100% from the pie chart. The largest category accounts for 58% of the data, followed by the blue category (40%). The last and smallest category therefore has 2% of the data.
The blue category is of the most interest, and the designer breaks that up into four sub-groups, three of which are roughly similarly popular.
The puzzle is the identities of these categories.
The sub-categories are directly labeled so these are easy for German speakers. From a handy online translator, these labels mean "definitely", "probably", "rather not", "definitely not". Well, that's not too helpful when we don't know what the survey question is.
According to our correspondent, the question should be "of those who plan to buy the new ticket, how many plan to use their car less often?"
I suppose the question is found above the column chart under the car icon. The translator dutifully outputs "Thus rarer (i.e. less) car use". There is no visual cue to let readers know we are supposed to read the right hand side as a single column. In fact, for this reader, I was reading horizontally from top to bottom.
Now, the two icons on the left and the middle of the top row should map to not buying and buying the ticket. The check mark and cross convey that message. But... what do these icons map to on the chart below? We get no clue.
In fact, the will-buy ticket group is the 40% blue category while the will-not group is the 58% light gray category.
What about the dark gray thin sector? Well, one needs to read the fine print. The footnote says "I don't know/ no response".
Since this group is small and uninformative, it's fine to push it into the footnote. However, the choice of a dark color, and placing it at the 12-o'clock angle of the pie chart run counter to de-emphasizing this category!
Another twitter user visually depicts the journey we take to understand this chart:
The structure of the data is revealed better with something like this:
The chart doesn't need this many colors but why not? It's summer.