Introduction
The Tableau Iron Viz Championship takes place during the annual Tableau Conference. There are three feeder contests and the first of these in 2017 was entitled ‘Iron Viz: Geospatial Contest‘. If someone had asked me a couple of weeks ago if I was going to enter, my answer would have been something along the lines of “Are you having a laugh, I’d be way out of my league”.
A few weeks ago I tuned into a ‘Data + Women’ virtual Tableau User Group which you can watch here. It was hosted by Chloe Tseng and Brit Cava. Chloe spoke about the contest and pointed out that people only ever remember the winners so it doesn’t matter how many votes you get. Lots of sound advice was dished out by the inspirational speakers and I’d recommend listening back. The below tips really resonated with me:
- Don’t wait until you’re 100% ready before taking on a project
- Put yourself out there and collaborate in the Tableau community
- Move out of your comfort zone
- Give yourself stretch goals
With these words still echoing in my ears I decided to submit an entry. I’m very glad I did as I’ve learned a lot in the process.
The Process
Wednesday – Demystifying spatial files
It was Wednesday evening and I was creating a viz for Makeover Monday. The topic was March Madness and I came across a cool viz by Shine Pulikathara on Twitter which used the same data-set (see data for Week 12). This led me to his blog about how he became an Iron Viz champ in 2015 and got me thinking. I started to read up about the competition and had a look at some of the entries already submitted. Adam Crahen had set the bar very high with his amazing viz but I tried not to let that scare me away. I also came across some other great entries already shared on Twitter by Ken Flerlage, Ravi Mistry and Pooja Gandhi.
I was curious about these mysterious spatial files. Luckily the kind people of Tableau Public had made it rather easy, with learning resources readily available. I watched a short video which told me that the most common type of spatial file is a shape file (.shp). These are normally downloaded within zipped files. You simply select ‘Spatial file’ when connecting, point to the .shp file and make sure the accompanying files are in the same folder.
The files have a ‘Geometry’ field which dictates how the locations are mapped. The values I came across were ‘Point’ (see below), ‘Polygon’ and ‘Multipolygon’ but there are many more. The data duo have a number of interesting blogs about merging shape files in qgis, nearest neighbour join and QGIS steps.
There was an informative blog by Kent Marten which supplied lots of examples of data sources. The first few data sources I experimented with came from the Irish Central Statistics Office. I visualised unemployment levels of people under 25 in Ireland in December 2016 (see below). A story wasn’t jumping out at me from the data I looked at and I called it a day. I hadn’t selected a topic but at the very least had successfully connected to shape files!
Thursday – More learning and picking a topic
The next day while scrolling through posts on Twitter I noticed that Sophie Sparkes was hosting a webinar about geo-spatial files. I registered and tuned in for 30 minutes to ensure I had the basics down. This is another useful resource if you want to learn more and explains how to join shape files with other data sources.
That day I also stumbled upon a tweet by Alberto Cairo about the current state of emergency in Yemen and Somalia:
The severity of the situation shocked me and I felt compelled to find out more information online. I often moan about trivial things (such as “grr, this file is taking more than 30 seconds to load” or “will my daughter ever just nap so I can finish my blog”) and reading something like this gives me a jolt and makes me appreciate what I have. I thought it would be valuable to educate myself about the situation and present this story in a visually engaging way. I had a topic for my entry.
Friday – Designing the layout, researching the topic & exploring data
I thought a long form dashboard split it into three panels might work for the layout. While my daughter was busy coloring, I robbed some of her paper and crayons and drew out a rough format of how I could structure the information. I would start with an overview of the global situation and then focus on Yemen and Somalia individually. I read several articles about the current high risk of famine. I noticed that some headlines focused in particular on Yemen and Somalia and decided to limit my analysis to those two countries.
I first set out to find data-sets that would help give the story context. HDX was one of the sources listed in Kent Marten’s blog. I found a data-set which listed all countries currently considered in crisis and contained various indicators for each country. Another source mentioned in the blog provided global and country level spatial data with a drop-down menu with various options (see below).
For the map of Yemen I downloaded the shape files with Administrative areas specific to Yemen. I then joined it to another data-set which provided a breakdown of the ‘people in need’ in Yemen. I found a number of files on HDX focusing on Somalia and settled for a data-set that showed the drought severity.
Saturday – Fine-tuning the analysis and formatting
The viz was taking shape so I decided to look for feedback from my family. When you are very focused on something it’s easy to develop blind spots. Things which seem obvious to you might not be to others. I subsequently modified headings, instructions, legends and Tooltips to ensure that the reader had sufficient information.
One suggestion was to show the total population of the countries in crisis so I decided to add an additional data-set that contained the population. I added this to the workbook, blended at a country level and calculated ‘people in need’ as a % of the total population. I decided not to clutter the views and simply included this information in the Tooltips.
Sunday – Finalising the text and adding finishing touches
I then spent time finalising the text in the text boxes, fine-tuning the formatting and validating the results. The result is by no means perfect but I achieved a few things I set out to do:
- use a long form dashboard which I hadn’t done previously
- make the story flow clearly from top to bottom. I wanted the key messages to be clear and legible if someone was glancing at the image on a smart phone
- endeavour to give clear instructions to people interacting with it
- include large summary figures
- use images of different people at the top and bottom to emphasise the message
See an image of the finished product below or click here for the interactive version on Tableau Public.
Thanks for reading. I welcome any comments or constructive feedback.
Getting vizzy with it 