Ryan Sleeper Visits Milwaukee

Ryan Sleeper is a celebrity in the Tableau universe. He is a Tableau Zen Master, Iron Viz Champion, and author of the Tableau Public 2015 Visualization of the Year. Add to that author of a new book called Practical Tableau (O'Reilly 2016). The release of his book brought Ryan by the Milwaukee Tableau User Group two weeks ago for the kickoff of his book tour and we were not disappointed.

We received a preview of the type of information laid out in his book. For example, he spoke about his Triple Crown Framework:
  1. Audience
    You should know who your audience is, what they are looking for, and their perspective. Are you presenting to clients, the CEO, or the public? Either way, your mom is a good litmus test, as every visualization should easy to understand without much explanation. It is worth understanding certain psychological concepts when attempting to display your data and Ryan lays it out in this post.

  2. Data
    Prepare your data before it gets to Tableau. Think about the format your data should be in for the type of visualization you want. For example, instead of having your dates going across the top of your file, in individual columns, have them in a single column. Ryan suggests using context filters when you're developing your visualizations, which contradicts what I've heard in the past. I suppose it might depend on the size of your data as a context filter creates a temporary table of a subset of your data.

    While Tableau currently has some data preparation abilities (with more coming in their Maestro product), they are extremely lightweight. I use Alteryx frequently to prep my data before it gets to Tableau and will be writing up a post soon about how it fits into our suite of tools as well as how my wonderful colleague has developed a system that allows us to automate our workflows using our enterprise scheduling system.

  3. Design
    This step is to find and communicate actionable insights. Ryan suggests (and I've already started to implement) reducing the saturation of color or making your visualization a little transparent as it's less harsh on the viewers eyes. Another big rule is less is more - keep it simple. The following gif explains Edward Tufte's rule of avoiding "chart junk". I tell the individuals I advise that unless a visual element adds to the understanding of your data, you don't need it.

Ryan went on to demonstrate to the group how to create a dumbbell chart. Essentially you have a dual axis of the same measure. The mark card for the first measure should have the mark type be a circle with the dimension field on the color shelf, while the mark card for the other measure should have the mark type be a line.

The visualization he used as an example is below. There is a toggle to normalize the data or have each game in the data set represented with the same starting point. There is a second toggle which allows the user to sort the chart in either chronological order or margin of victory. You can play with the dashboard live on Ryan's website.

I will leave you with this random thought - whether you look at Ryan's website, Twitter, or dashboards he is consistently using the same color scheme. I appreciate that immensely and have yet another project to add to my list for the year. What do you think of this?

The #MKETUG has three other meetings currently scheduled for the year, which I wrote about in this blog post and which you can register for on the right side of my website. Hope to see you there!