« Mid-week light entertainment: The Economist imitates USA Today | Main | Rotating circle, loose ends, on-line dashboards and charts »



I'm definitely in favour of the side-by-side boxplots for this.

Principally because the horizontal-scale variable "round" is discrete (ordered categorical if you wish). The takeaway is that there is a downward trend in mean/median, but an awful lot of variability (that decreases slightly as round # increases). This is exactly what a boxplot (or series of boxplots) shows, and it does so without clutter.

The top plot has introduced a spurious horizontal scale (unless there is meaning to left side vs. right side of the Round 1 box in the first plot, which is not mentioned). The second plot has those same issues (what is the horizontal scale? Is it pick # overall?).

If the statistical issue in question is "do players picked in an earlier round of the draft tend to have a higher value?", boxplots will answer that; your plots appear to go after the related but different question of "is player value related to draft pick number", with the round # of the draft being incidental to that. For that, a lowess curve on plot #1 would do the job for me.


"Our eyes cannot judge density properly especially in the presence of over-plotting."

Make the dots transparent?

The comments to this entry are closed.

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