Analyzing Literacy’s Impact on Synapse Formation AKA Making Brains in Tableau

I’ve always been moderately obsessed with the brain. For starters, it’s creepy looking. I remember a cartoon where the villain was a brain in a storage device and, well, ever since then, I’ve wanted to know more.

So, when the latest IronViz feeder focused on books, brains were the last thing on my mind. No, no, I planned to document my library, which would definitely showcase my love of anything ‘neuro’-related.

Except, my web site got hacked and that back-burnered everything. When I finally got some time, I realized I was down to about a week and a half. So, I went looking around for the usual suspects, linguistic analysis. Unfortunately, I started some work and realized the data I really wanted was behind a massive, massive ($2K+) paywall. So, there goes that idea…

I looked into library funding and several other avenues. I tried a desperate scan of my books, but realized to do anything really interesting, I’d need more time. So, how did I get to brains? I gave up, for starters. Then, I had a few chats with people and discovered this gem:

I also ran into a few discussions on Twitter about literacy. What if I took my neuro-obsession and made it a viz? This acted as a great test to see what I really knew (a lot less than I think I did) and challenged me to figure out polygons yet again…there’s still more work to do in this area. Keep in mind, I had this epiphany on Friday and the deadline was Sunday. Secretly, my brain is the afore-mentioned villain.

The most essential piece of my viz would be the brain. I did not want a static map and I wanted to get fairly granular in the brain. I needed a way to get polygons. Now brains are not like states or stores. Brains are wrinkly, curvy and terrible in all the ways possible for hand-drawing shapes to any decent level (AKA, using my mouse with all the hair in the laser to make Tableau Toddler art). Lots of googling told me SVG files were the way to go. Surely, someone had a good SVG file of the brain. Except, not really. Most files I encountered were pretty flat and didn’t capture the brain in a way that was interesting to me.

So, I found some SVG editors and made good friends with the SVG files from Wikipedia. I ended up shelling out $10 for Boxy SVG on my mac, honestly, because it had the best description and didn’t require a whole lot from me. You see, SVG is yet another file format that renders as XML. Here’s the short list I have of things that also use XML:

• KML
• Mind mapping programs
• Alteryx workflows
• Tableau!

XML matters because it shares and passes data. I like data.

Those names and points pass with the file. Now, I needed an intermediary to translate the XML to a tidy point path. On to more googling and asking people. Ken Flerlage helped me out and ran it through his python script.

With the brain in Tableau, I started experimenting.

Now, the brain is not a straight-forward organ. When we view it in 2D space, we miss a lot of it, so we view it in slices. I had 2 of the 5 types of slices. It has a number of ways of breaking it down into areas and locations. I also could not find a good synthesized resource here, so it’s added to my list of future projects. My view is also only part of the brain. There’s also the cerebellum (old brain) and the brain stem (this is more important than it sounds). We often focus on the cerebrum, which is where some argue the ‘mind’ exists and what’s pictured above.

Brain technicalities sorted, I still didn’t have a viz. I had brains, but no story. I wanted my users to interact with the brain, to touch it, and explore it. How often do you get to do that? I started with my study guide (above) and looked around.

We often see the brain in context of MRIs or brain activity.

This provided a good lead, but meant making more shapes/points…I decided to use PowerTools Custom mapping. A word to the wise: when you mix and match methods, they don’t always align. I had to solve for that with a calculation. Tableau does not have density plots (yet) so I needed a different way to create that. A quick LOD found the median Y axis and compared the absolute difference from that. Combine that with a good color palette and we get this:

I continued making points. Because I had the mix of SVG files, PNG, and multiple methods of getting points, they didn’t play well. I opted for simplicity out of time’s sake and didn’t display the medial cut of the brain.

I also knew I’d go long-form “scrolly-telling” to help guide the user. There’s far more text on this than I normally put and less interactivity than in a normal viz. I also tried to keep it to a width I felt would work on tablet. I went dark in part to mask imperfections in the polygons and also to keep the emphasis on the brain activity or MRI effect. The font, Brainy, made me laugh, so it got used. I used various illustrations from The Noun Project and tamped them down to keep the focus on the brain.

One of the key points I wanted to get across is the idea of neuroplasticity. The brain changes how it operates based on new information. Preliterate people see written letters and numbers differently than someone who can read because of how the brain works.

Lowercase ‘b’ and ‘d’ make great examples. They’re mirror images of each other most of the time. If this happens in nature, we usually consider it to be the same object. Letters are different and it takes A LOT of time to teach our brains this and I really wanted the literacy chart to be a part of this viz. I emulated it and included a table in the legend.

I started framing this out in a dashboard, trying to ensure the message was clear. Lots of feedback went into making something like neuroscience accessible in a visualization. I went back and forth between my data sources, adding in information, changing things, and so forth.

In the final piece, you can see how how so many parts of the brain interact with reading. To me, it’s astounding.

This definitely gives me more to explore. Like Marvel, there’s some fun easter eggs if you explore and a bonus with the credits.