VizGames 1: 5 Stars in Yelp Data

The joke in interpreting is if you put 5 interpreters side-by-side, you’ll get 5 interpretations of the same utterance. In reality, that’s all very true. One game we played frequently to improve our skills was that of renditions. We’d try interpreting something to different segments, seeing how things like the setting, audience, and even occasion influenced our take.

Now, in CrossFit, you have annual challenges called Games. Basically, it’s a way to push to the max and find the fittest of the fittest. Folks, I assure you – you won’t find me winning that ANYTIME soon.

What I did want to do was blend both of these into a sort of ‘games’ meets ‘renditions’: I’m proud to launch the VizGames.

Now, let’s not get delusions of grandeur here. VizGames take a concept and try to find the many ways to solve it. But, here’s the thing: you can play, too.

Our current challenge is brought to you by Brian Smith. He attended the conference and wanted to play around with visually showing 5 stars of the Yelp data set used the the Tableau Advanced Calculations class. He preferred to do this via calculation. We chatted about this in the airport and I dedicate this #VizGame to him.

 

Solution 1: Index with bin in columns displaying average of stars.

This seemed to best match Brian’s description of what he wanted to achieve. I liked it the best and added in some other info via tooltips.

 

Solution 2: Index with bin in columns displaying all stars + average.

This is pretty much the same solution as 1, but with the info in tooltips displayed in the graph. The average is the larger star, but it seemed to get lost a bit when the average (ex: 4) came from different scores (3 and 5) but was a value not selected by Yelpers (4).

 

Solution 3: Shape files for each value.

This seemed logically to be the easiest to reach, but does not execute well in this application. I think it has more to do with my photo editing skills (minimal) and lack of robust editing program than the solution itself.

 

Solution 4: Bars plus cutout image.

I liked the theory of this prior to applying it. If I were comparing a very small subset of variables (ex: 2-3), this would be viable as we’ve seen with other cutout + bar visualizations in the past.

Wanna play? Download and come up with your own solution. Comment on the blog or tweet @windscogley with #vizgames with a link to your solution(s) on Tableau Public.