Week 6: making graphs

I have made some progress with R, mostly because of one particularly helpful site called “Cookbook for R” that had good examples of how to make plots in R.  In particular I had been struggling with making a plot that showed the number of similar observations.  I could easily graph each individual observation independently, but it didn’t tell me anything new.  I could also graph the number of similar observations as a histogram.  This was somewhat useful but what I really wanted to produce was a line graph that showed me the distribution of my data.  My PI showed me a page on Cookbook for R that had detailed examples of how to produce the graphs that I wanted.  It was easy to modify the code to suit my data.  There were even options for making the graphs prettier or adding keys and lines to show the means.

The graphs allowed us to see some interesting aspects of our data.  Visually, we thought that Asclepias exaltata leaves looked narrower and longer than Asclepias syriaca leaves.  But, when we compared the ratio of leaf length to leaf width in each species, there was no significant difference.  We did see, however, that the apical angles and the basal angles for each species were quite different from each other.  So far, we haven’t seen any super clear indication of hybrid populations, but there are some outliers from each species that fall more in the other species’ distributions.  These will be good specimens to go back to and take a closer look at.

Also, I succeeded in running an example of code for doing a PCA analysis on the iris dataset.  I was then able to rework the code slightly and run it for my own data.  I feel like I sort of understand the background behind principal component analysis but my brain can only comprehend it for a limited number of variables.  Once there are more than four or  five axes, I don’t have the processing power to picture what is going on.

PCA1.2