« Flooding the Himalayas | Main | Look what I found: two amazing charts »


Paul D

Hi there,

I'm an occasional JMP user, and I've tried rotating the y-groups and I can't see anything obvious, short of doing it manually in Inkscape or [insert editing program of choice].

One thing I can do to improve the graph though, is arrange your regions top to bottom by average happiness, as per the small multiples chart:

You can do this by selecting Region Column > Col Info > Col Properties > Value ordering

Finally, I like the scatterplot analysis. Again changing the order of the plots, I can see that NA / Europe / LA are happier and don't correlate with each other. Sub-Sahara is sadder and doesn't correlate with the others. The 5 remaining middle distributions are highly correlated and could be treated as a dingle region in any further analysis / explanations


I noticed that in your scatterplot matrix (which, by the way, looks great) you followed the convention of only using the lower triangle. I'm wondering if this is because you think this is the way plots like this ought to be presented. Personally I find it quite an annoying convention. I mean, I get the idea--it's a symmetric matrix, so half of the elements are redundant--but visually speaking, it's just a lot easier to get a sense of the relationships with a particular variable by scanning across the row and/or column associated with that variable. With one triangle removed, not only do you have to change the "direction" of your scan midway through, but the axis on which the variable of interest lies switches sides!

Bob K

How come you prefer JMP over R for some of the graphs?


If you hover over an axis in JMP until you see a hand (Grabber tool), double click to open the axis specification window. There you will find a section called Tick Label Orientation, with a drop-down menu including options such as horizontal, vertical, and angled.


Laura: thanks, I'll check that out.

Bob K: I wrote a post about the Graph Builder function in JMP some time ago. It's like a sandbox that you can use to explore different chart types. R is great if you already know what type of chart to use-otherwise, you throw away a lot of code. R is great if you want to control tiny details such as using irregularly labeled axes, but if you're just making a standard chart, why not save some time?

The comments to this entry are closed.


Link to Principal Analytics Prep

See our curriculum, instructors. Apply.
Kaiser Fung. Business analytics and data visualization expert. Author and Speaker.
Visit my website. Follow my Twitter. See my articles at Daily Beast, 538, HBR.

See my Youtube and Flickr.

Book Blog

Link to junkcharts

Graphics design by Amanda Lee

The Read

Keep in Touch

follow me on Twitter