Seeing fast is sometimes too slow

Humans have a remarkable ability to connect dots, extrapolate lines and build mental images of the world through vision. Faster than thought, most of the time.

Those features aren’t just ancient evolutionary facts from millions of years ago, they still work today. Sometimes the images appear fast, sometimes it takes a moment or two to have the brain connect.

This, spotted on a t-shirt. Took my brain a little too long to “see” image and ask for a photo.

Can you spot the animal that isn’t really there?

100% Text Charts – In Action

Data viz turns abstract data into tangible insights. In my past post I shared an abstract how to make a 100% bar chart with text. Now, I’ve gone and made it real for you. [scroll down, its a ways to the bottom]

          

Check out the first tab, where I neatly line up a few examples for you. I took the bar chart idea and thought, what else could we do. Well, why not a Lollipop chart, or variation of. I reckon there are more neat ways out there – hope you will try some.

I popped the view onto a map, second tab, like a geographic overlay. Useful for things the same size, and if you don’t want to fiddle with charts a lot. This can be useful on maps and more.

And best of all, the steps behind the trick. You can see in formula windows the way to extract parts of the text character, then sneak into the text settings on the marks tab to change the colour. Once the colour is changed and the text strings lined up, viola, you have it.

Hope I’ve been able to make the abstract text to chart idea into a tangible for you. Download the workbook if you want to see and try the magic yourself.

So I entered the Iron Viz challenge

I did it. After years of doing vis for fun and for work, I finally took the plunge to do one for the Iron Viz competition.

What the heck is the Iron Viz? Well, in the Tableau community, for some time, there is an annual competition to see who can craft the best viz, on stage, in 20 minutes. It’s modelled a bit after a cooking show competion, and there are some wildly interesting things that go into the various Vizes. There are also, if I’m honest, some entries I’ve seen that I’m sure I could do. Sooooo, this year I did.

The challenges were many.

Finding Focus Time: October is a long month, but somehow packed with events, work events, family events, get out and meet some people events, the regular weekly events that already consume a lot of time and, if I’m totally honest, some streaming events on Disney plus to catch. Oh, and I work like so many do, commute to the office too much for questionable reasons, have two kids to take care of, haven’t given up on regular exercise or getting enough sleep to be functional.

Choosing a Topic: I started trolling my brain for fun topics once I heard about the challenge, and saw its open to those in Singapore this year (previous years it was not, which slowed me down a lot). One topic I’ve had on the mind to viz is, as my tittle might be, “The Rise of the Bullies”. I figured some democracy rating data, some look over time and certain less savoury world leaders, some line and gantt charts and there might be a story. But no. So it mulled till late in the month, when I next thought about using the Economist’s Burger Index as inspiration to compare inflation. The data was there, I was sure, the idea approachable. Onto the quest for …

Useable Data: With a topic in mind, I could get the data for the Big Macs, and thanks to the awesome Economist index (full disclaimer: I am a passionate, multi-decade reader of the Economist) which posted not only the latest online, but data going back more than 20 years. Oh yeah. But then the inflation side, how to get that. Why do some countries have it twice a year, why can’t I find a set that lines up the timing? Why does a great dataset from the IMF miss out on certain countries that both have Economist data and lots of McDonalds to make the data better. Let’s just stay, and you may know the feeling, some data wrangling occured. Then…..

My Data story went wrong: As I pulled the freshly wrangled data together, had some cool text snippets and charts from varied sources getting ready to tart up, as the data got visualized, the story failed to materialise. I had a bunch of LEGO like pieces of a story to go, but the data did not support the story. It wasn’t bad, but it was far far far from being a good story to tell. And given I have a thing about accurate data, I could not post something about how BigMacs can tell you about inflation for about 2 countries, on about 1 period of time.

The clock still ticked however, the month end submission time approached and my stubborn “I can do this” habit remained strong.

In the end I went to some work from earlier this year that has a more compelling story, an easier dataset to work with, and looked not too bad to start with. I’ll save the detailed mechanics on that vis (since it uses some good Vizology techniques) for another blog. Suffice it to say, I made it better. I made it tell a better story, it got through my checks and I took the plunge to put it out there, to fill in the entry form and submit a Viz for some mysterious judges to evaluate (scary) and maybe the world to see if it qualifies or moves further along the process (also scary).

What did I learn, to share with you?

Start Early: Always keep some topics in mind to explore, try things out and keep work even that is in draft for a while. A month, where you are busy, flies by so fast.

Practice: Being able to work with charts, visuals and colour combos quickly is important, not only at the last minute. It lets you pivot well if the data story is nor materialising, and to assemble a good looking starting point to polish.

You have to try: No one ever won a competition they didn’t enter. There is no way I’d ever get anywhere in Iron Viz, or other challenges, without the step to at least enter. We are all our fiercest critics and that vis you think might be just ok, is likely magical for many others. Have confidence, make the effort and enter.

Keep Learning. Keep Trying

ps: Shortly after this post I learned to Read and re-read all the rules and FAQs. Alas I missed out there being a particular theme to the contest (in my defence, the theme is down in the FAQs, not in the general descriptions) , and my story was off the mark. My bad, but learning lesson.

Castle Charts and Height Inflation

I recently enjoyed a trip back to Japan for some great food, family fun and the occasional historical adventure around the Tokyo area. We stopped at one place for some history, Odawara Castle, in Kanagawa. This castle is often on the list of things to see, partly as it’s a short Shinkansen ride away from Tokyo and , as I learned, it’s fairly historic and one of the tallest in Japan. History wise, it played a role in the 1500s consolidation of Japan being the last major castle to surrender to Toyotomi Hideyoshi. Height wise, as advertised, it’s in the top 10, number 7 in fact. [total side note, nice to know I’ve been to 4 of the 5 highest castles]

Yet the viz eye in me spotted a small problem. Overall it’s a good chart representation- scale at the side but without much overbearing numbering. Good visuals of the castles to attract the eye and a precise and consistent height out to one decimal place for all. The lines on the Y axes are not over bearing, and use gradient and size well between the 10m and 5m increments. All good so far.

Castle Chart

But look closer, really close, at the Y axis lines, for the star castle, Odawara. Listed at a 28.7m height. Doesn’t that very top top of the castle look to be over the 30m line? Pretty close between position 6, 7 and 8 on the charts, yet there does appear to be a bit of Castle Height Inflation going on.

What do you think? Should they redo some graphic sizing to be Viz chart accurate with the axes and all? Or leave it be – it’s telling a story and a little artistic license surely is allowed in your Castle Chart.