The graph and the serpent: a cautionary tale

The other day, my friends were walking out to the bar, which is right on the beach about 200m away (don’t be too charmed, its a sandy but really high-energy beach.  Not inviting at all.)  I really wanted to go too, but slightly stronger was my obsession with a figure that I almost had organized correctly.

Finally I got my figure! I had to learn some new stuff to do it.  A while ago I pledged I would only graph with ggplot2 until I understood it.  So that took a while.  I also read a wee bit about ordination techniques and spent some time selecting a pretty colour scheme.  The result!


The arrows show the before->after trajectory during the experiment, and the dots show the original, unmanipulated communities (more info on the experiment later!). Each panel is a block.  Yes, we named our experimental communities after Dr Who actors.

Lovely, isn’t it!? And a proper thing for an ecologist to be doing with his time in the field — already graphing and analyzing!  I felt pretty good about myself.

Then I caught up with my friends at the bar — and they showed me this photo:

I had missed seeing this awesome snake eating this lizard! They say that when then approached it dropped its prey and moved away, but when they saw it posed no threat it returned to swallowing the (very dead) animal. Photo Credit: Pedro Trasmonte (not me!)

So remember, ecologists — graphing is fine, and earning computer skills is awesome.  Just make sure you find time to look at awesome animals, too.

4 thoughts on “The graph and the serpent: a cautionary tale

  1. I seriously LOVE that your experimental plots are named after doctor who actors! do you have 11 plots? or just chose the 5 best ones?

  2. I also am a big fan of your experimental site’s names. Do the site’s characteristics somehow align with the traits of the their respective Dr Who incarnations, I wonder? You will be nerd heroes if you use these names in the publication (or at least a hero to this nerd)! Graphs are looking pretty nice. What ordination technique did you use?

  3. Loving the plots Andrew.

    One improvement needed is to have the axes scaled equally so that a 1 unit change in the x-axis direction is the same distance on the diagram as a 1 unit change in the y-axis; distances on these plots mean something and should be preserved as a far as possible. Luckily this is very easy with ggplot, just stick + coord_fixed() on to the end of the call that creates the plot.

    1. Great! thanks for the advice! These look prettier now — I’ll post the latest version with your suggestion. I can see I’ll have a bit of homework to do before I can contribute to `ggvegan` (which I forked the other day)

Comments are closed.