Software

Quantitative Tools for Variance Explicit Ecology and Herbivory
Using herbivar to identify the statistical distribution of herbivory observed in a lab experiment deviates strongly from neutral expectations and an herbivory distribution observed in the field.

I write a lot of esoteric code for my research and sometimes stumble across small statistical methodological ideas that are not publishable. Because these things are sometimes useful to other people, I’ve made them available in my R package.

I actively maintain the herbivar package available on GitHub. This package makes several methods used for analyzing variability in ecological data available. Currently, it supports fitting a neutral herbivory model to data using Maximum Likelihood Estimation (MLE). It also contains several helper functions for comparing distributions and assessing model fit. In addition, using the imager infrastructure, the package also supports automatic processing of leaf images, mostly in the context of leaf scans, that can retrieve rich within-leaf herbivory data. The package is integrated with the spatstat packages that allows for spatial analysis of these data.