Hi All,
Here are a few things I'm looking into at the moment:
Potential to use Bayesplot for analyzing success of Bayesian models.
Start implementing a training/ testing split in the dataset (maybe use the dataPreparation package).
Refine GP fit (BRMS) parameters and default STAN priors
Wrap code into chunks and create the infrastructure needed to easily build and test models for various species and with various model parameters.
And a question... do we need to scale the data before using it in the BRMS model?
Notes on projects and initial reading https://docs.google.com/document/d/1JJJ-5pJGam7dYTzAAjHBdEZkHYOVlkc7GdZJk8KfzTU/edit
Prototype analysis using brms package https://github.com/HuckleyLab/SDMpriors/blob/master/SDMpriors_brms.R