Tuesday, August 12, 2014

My first interactive visualization with Tableau

Classes at the University of Miami begin two weeks from now, so I'm quite busy getting syllabi, lecture slides, and projects ready. Between the new UM Visualization website and the Places&Spaces exhibit and lecture series, I haven't been able to sit down and start planning for lessons until today. Oh, I've also been writing that book that should be published at the end of 2015, of course. More news about that soon, hopefully.

This Fall semester I'm teaching a new version of my CVJ522 Infographics and Data Visualization grad course (read an unedited draft of the syllabus,) in which I'll have students from both our MFA in Interactive Media and MA in Journalism programs. Therefore, I cannot just focus on news visualizations or print infographics. Instead, I decided to give students the opportunity to choose the tools and styles they prefer. Some of them will come to class knowing a bit of Processing and even d3.js, for instance.

Other students presumably won't know how to use any visualization design tool. To those, I'm recommending Adobe Illustrator (I'll use my own video tutorial,) open source tools like Lyra, Datavisu.al, BlockspringPlot.ly, etc.

And Tableau. That has been my own elephant in the room this summer. The challenge I was facing was that, even if I had taken a look at the program in the past, I had never really created anything real with it.

So today I opened Ben Jones' Communicating Data with Tableau, took a look at some free video tutorials, and started playing with a small dataset that I had gathered months ago about poverty, income, and race. In just four hours —counting the time I spent reading and watching a bit,— I produced the visualization below. I feel that I've just scratched the surface of this tool. If your average Jane or Joe can do something like this in half a day, imagine what can be accomplished after a week or two of serious tinkering.

Facebook's Andy Kriebel and Tableau's Ben Jones himself helped me with the last portion, which involved merging two connected scatter plots. Geeky stuff. Thanks, guys.

IMPORTANT NOTE: Remember that this is just an exercise. I didn't double-check my spreadsheet or tried to analyze the numbers to extract stories from them. There are also some errors and inconsistencies, like the fact that median income is used in the first chart, and average income in the scatter plot. I was just playing with the software.

UPDATE 08/14: Jeffrey Shaffer has improved the visualization. See second version after the one that I designed, or here. Quoting: “Just for fun, I added a play control with history to your second chart (only work in Tableau Desktop) and a label for the year by tricking it with a dummy point in the middle of the chart.”







5 comments:

  1. Great work and a fantastic example of what someone can do in Tableau is a very short amount of time!

    ReplyDelete
  2. I've been using it for about seven years, and I never thought to do a time series on a scatter plot. What a great idea; I'm going to try it out on some data I have.

    ReplyDelete
  3. Connected scatter plots can be useful sometimes (not always). See these posts

    http://www.thefunctionalart.com/2012/09/in-praise-of-connected-scatter-plots.html

    http://dataremixed.com/2013/05/connected-scatterplots-and-strikeout/

    ReplyDelete
  4. Nice work Alberto - amazing to see how far you've come in two days!

    ReplyDelete