It’s been a while since I’ve done one of these, and I wasn’t certain which viz to look at. Step in Ken with a decent suggestion:
Athan is an IronViz Europe finalist, so he has pedigree – and also humility:
- Explanation in the header and conclusion in the footer – a storytelling structure
- Good use of colour
- Effective depiction of seasonality
I’m a firm believer that the best visualisations aren’t always the ones with complicated calculations and intricate floating of countless objects. Effective storytelling is clear and concise, so there’s plenty that can be learned here.
Athan was correct in stating that it is an easy to build chart, with the basic structure as shown below:
It was always going to be like this. We need Longitude and Latitude for the maps, and we need Discrete blue pills to break the view up with Year on Columns and Month on Rows. Why Discrete? Continuous causes this mess:
Something I failed to do in my own submission with this dataset was aggregate the measures to form the AQI Index, which Athan did like this:
Simple enough, and it could be even simpler:
When you have nice, discrete aggregations like this, then your Colour palette is controlled and you don’t have to be concerned about the ranges between outliers, such as you do with continuous measures on a diverging scale. This meant that Athan’s colour choices were simple and logical:
Andy’s recap mentioned the simple but effective tooltips, which look like this:
And are constructed like this:
So there’s a sneaky series of tooltip calcs there, with the express purpose of enabling you to tie the colour palette used in the viz through to the tooltips. Those calcs look like this:
So they basically mirror the AQI calculations shown earlier, but they have to be created independently in order to allow you to colour each possible outcome in line with the overall colour scheme. If I just put the AQI calculation in the tooltip, I’m denied that flexibility.
My beady eye spotted one more thing. I noticed on the worksheet that the headers were shaded, which is something that I didn’t ever think to explore before:
There’s so much functionality in Tableau that I still pick up these tiny details pretty much every time I load up the software. Why has this had to be used? If I take the shading off…..
The dashboard itself is shaded grey, so the header stands out without shading. And that’s pretty much a wrap. The final dashboard itself is entirely tiled, as demonstrated by these images from the 10.4 Beta in which this deconstruction occurred:
Thanks Athan for a great viz, and Ken for a great suggestion!