Today’s dataset was a small one. The only way I could find an interesting story was to group the brands by their respective manufacturer. It then became clear that Cadbury is losing its popularity. No fancy graphics this week. I played around with sweet wrappers, but it looked cluttered, so I went back to nice coloured oblongs. I did try colour-coding by manufacturer and brand – but I got told it was too confusing – so it’s ended up simple.
I’ve decided to stick with “small blog” sized vizzes for a while – just for the fun of it.
A cute one, this. I wanted to use the metaphor of a race to illustrate the changing positions during a Formula 1 championship season.
It was a nice dataset for analysis. However, I did have to do a bit of tweaking to get my idea to work. Simply put, I needed Tableau to calculate a running sum, so that I had a total number of points for the season so far for each driver. This was fine – except I also wanted to show the constructor. This would be easy if there was a one-to-one relationship from driver to constructor in a season. Unfortunately for me, drivers do change teams mid-season. And whilst Tableau now has the capability to do “Level of Detail” calculations, it can’t do them on a running total.
Ideally I would have wanted a cumulative snapshot type of dataset because of the way Tableau works. However, that’s a lot of effort for what was meant to be a quick viz. So, I’ve made some simple changes to the dataset instead. I chose to only show the constructor that the driver was driving for at the end of the season. This involved adding a column to the dataset to record the driver’s final team. I also added a dummy record (round 0) for each driver in a season, so that I could start the line at 0 for everybody.
The other problem is that the pages shelf doesn’t work on Tableau Public (or Tableau Server, come to mention it). Here’s a gif to show how it works, as well as the interactive version.