Python code to summarize morphospace/ niche diversity, tissue availability, and proportion of specie threatened/endangered across all bird species by order and by family.
The script first does a discriminant analysis of principal components using scikit-learn using morphospace and beak shape PCs from Pigot et al. 2020. It also does a classification using 80% test data and will spit out classification accuracy to stdout. Typically ~65% for classifying to order from morpho/beak PCs.
Then it parses at order/ family level using data from a bird taxonomy database joined with the Pigot et al. dataset and a large VertNet query database containing tissue accessions for bird species to calculate various neat things and make some plots.
This script parses a large tab-delimited database of all tissues accessioned in VertNet. To access the file of all bird tissues in VertNet, go here: https://osf.io/d9szb/. It was too large to host on GitHub...
The goal of this script is to select a minimum number of museums required to find X number of tissue specimens for each taxon in a given family. Settings for the family to search and how many specimens are set at the beginning of the script (open it in a text editor).