In Search of Fame and Glory in Las Vegas: Visualising the Periodic Table using Wikipedia data

Hi Folks,

For this last week i’ve been on a quest to seek fame and glory (or at least a free trip to a massive party) as an entrant in the TCC15 Las Vegas IronViz Championship.

You see, last September I was lucky enough to be in Seattle to watch the 2014 Ironviz contest at the Tableau Customer Conference.  It was awesome! Three contestants, up on the stage, big screens behind each of them; with only 20 minutes to wow and amaze an audience of around 1000 people by creating a professional looking dashboard that was judged right there on the spot.  All three were using the same Yelp dataset of Las Vegas restaurant ratings.  The show was really amazing – they all came up with some pretty cool stuff – and all whilst having a whole lot of fun. The 3 contestants qualified for the final by winning one of 3 feeder contests held during the year.

This year, I’d really love to be up on stage as one of the 3 contestants.  The first feeder comp is on right now (only just found out last weekend!) The contest rules are quite simple: “Create a visualisation of any topic you want: as long as the data was found on Wikipedia”. 

So – needless to say this last 7 days have seen me diving in deep to Wikipedia and scraping out a great dataset to visualise.

I knew my job was going to be much easier if I could find a nice big fat table of data.  So, after doing a google search for ‘Wikipedia table’ I quickly came across many results for ‘Periodic Table’.  After just a few minutes of clicking around I could see the potential of the various different ways of slicing and dicing and measuring the 119 different chemical elements.  I also thought that this kind of information might have universal appeal to anyone who has ever done high any school chemistry before. (i.e. craploads of people)

Next, I thought “Better make sure I’m not reinventing the wheel”.  So, a quick search of Tableau Public vizzes for ‘Periodic Table’ quickly revealed there was already a viz by Ben Jones aka Data Remixed which made use of Wikipedia Periodic Tableau data. I really enjoyed the witty blog, and thought that It was a good first pass, but in the spirit of tableau public – I said to myself “I can take what Ben has done, build on this and make it a whole lot better”.  I know he wouldn’t mind.

So in-between watching World Cup Cricket games (its quarter-final week – Go Aussie!); re-crawling data for my cricket vizzes, upgrading a client from Tableau Server v7 (I’d forgotten how ugly those grey shaded dashboard sub-titles were); and the occasional Clash of Clans raid – I set about putting something together.

I spent a fair bit of time exploring Wikipedia, Wow – what a lot of info!  I not only tapped into one big fat table of data, but managed tap into 8 separate tables of wiki data! – either by web scraping using import.io, or via a simple copy and paste.

Joining the tables all up using elemental symbol, or elemental number as a key was quite a simple job – and produced a nice rich data set – that might even be of use to someone someday – quite aside from my viz.

In the end I found I was able to set up a slick looking elemental periodic table and to have my interactive audience slice and dice the elements using some cool colours in 9 different ways. (10 if you count none as one of them 🙂 )

and measure them in 11 different ways:

I had to reach deep into my bag of Tableau consulting tricks to get a single dashboard that did all I wanted it to.  Some of my tricks included:

• Web URL image actions.
• Using a parameter to hide/show various floating windows.
• Background maps (check out % of the Human Body Mass measure).
• Using a table calculation of size() with actions to hide and show floating windows.
• Massive amounts of string calculations to cleanse and transform the raw data. (What do you do when tableau treats 260×10−9 as a string!?!)
• Some really nice colour palettes.
• A jitter index calc to fatten out my strip plot.
• Parameter driven Metadata information tooltip hovers.
• Tableau calculation ranking filtering and sorting.
• Action filter based key word search and highlight tool.

I even had time to shop my alpha and my beta releases around to a few folks for UX comments and some great ideas along the way.   Thanks guys!

So anyway – without further ado – here is a link to the finished product.  I hope you like!!! – and I hope you can vote me up when voting starts next Monday. (i’ll add a link via comment below after voting starts)

cheers

Craig