Episode 27: [Mini] 3 Things I’ll Try To Do Better After Reading Tufte’s Book

 
IMG_0009.jpg
 

Welcome to episode 27 of Data Viz Today. I loved reading Edward Tufte's “The Visual Display of Quantitative Information” for the first time as part of the Data Vis Book Club! In this episode, I talk about the 3 things that I'll try to do better in my data viz now that I've read Tufte's book. What about you?

Listen on Apple Podcasts, Google Play, Google PodcastsStitcher, SoundCloud & Spotify.

  • Welcome! I'm Alli Torban.

  • 00:30 - I just finished “The Visual Display of Quantitative Information” by Edward Tufte as part of the Data Vis Book Club, and I wanted to share the 3 things that I’m going to try to do better after reading this Tufte book.

  • 01:40 - #1 First thing I want to try to better is to refocus on showing the data. The proportions, the variation, the context. It’s easy to get caught up in the design and trying this or that color or technique, but throughout the book, Tufte always comes back to his main point - focus on showing the data clearly - show data variation, not design variation. Try to erase as much non-data ink as possible, that’s the ink you use to print things that are not essential to the data information. You want to strive toward erasing as much non-data ink as you can. It’s a really valuable perspective to think in terms of data ink and non-data ink, and assess what’s truly important to communicating your data and insights.

  • 02:40 - #2 The second thing is that Tufte came up with some interesting techniques in support of minimizing non-data ink, which I want to try:

 White grid lines through data.

White grid lines through data.

One technique is erasing as many gridlines as possible and if some gridlines are necessary, instead of using black lines in the background, you could use white lines that just go through the data so it gives you a way to visually break up segments without adding any non-data ink.

 Range frame.

Range frame.

Another technique he calls the range frame, which is where you draw your x or y-axis, but only in the range where you have data. Like if you have a scatter plot that has data only between the values of 10 and 15 on the x-axis, then only draw the x-axis line in the range between 10 to 15 not like 0 to 15 like you normally would. It might be a nice minimal way to show the axes, and the reader can quickly see the range of the data points.

The book has lots of other little techniques like that helped to get me thinking about creative ways to reduce non-data ink.

  • 03:55 - #3 Third thing I want to try to do better, is not rely on color so much

He has a great section about not turning your data viz into a puzzle. He says that a sure sign that you created a puzzle is if the graphic has to be interpreted through a verbal process, rather than a visual process. Like if your reader has to rely on reading a legend and repeating phrases to themselves to decode what you’re showing, then you probably created a puzzle.

Tufte notes that often color creates graphical puzzles because we attempt to give color an order and assign it meaning, which means someone needs to decode it, so it can very easily get complex. So instead, think about using shades of grey because it has a natural visual hierarchy so there’s less decoding that needs to happen.

I think I rely on color a lot and go straight to it to start using it to encode attributes of my data, but Tufte makes a great case to keep things simple and consider whether you’re making a puzzle out of your graphics, and oftentimes color can be the culprit.

  • 05:00 - My final takeaway is that reading Tufte’s book “The Visual Display of Quantitative Information” gave me a refreshed perspective on data viz and was a great reminder to keep things simple and show the data, erase as much non-data ink as possible, and can be creative in the ways that you do this by not just erasing but using different techniques, and finally be weary of creating a graphical puzzle especially with color. You want to show data proportions, variations and context so your reader can gain insights, and as Tufte says, graphical elegance is often found in simplicity of design and complexity of data.

  • 05:45 - I highly recommend reading his book, not necessarily as a data viz rulebook, but as a valuable perspective to consider.

  • 05:55 - Did you read the book? What was your impression of it? Let me know on Twitter or Instagram.

  • 06:05 - I put together a Resources page with my favorite books, blogs and tools!


Allison Torbanmini, tufte