Files for analysing spatio-temporal plant species presence trends using presence-only data.
The idea of this project is to analyse the spatial and temporal trends of rare and scarce plant species observations gathered by Thames Valley Environmental Records Centre (TVERC). The data are presence-only records, meaning that there is no direct information on species absence locations. This forbids the direct inference of species observations probabilites using e.g. logistic regression. Instead, we build a Bayesian additive generative model that allows us to extract relative changes in the local rare species observation probabilities in comparison to relative changes for a selection of widespread "baseline" species. We assume a Gaussian prior structure which allows the use of the Integrated Nested Laplace Approximation (INLA) method for fast approximate Bayesian inference.
This is only a simple example code. Not all models and analyses are included, nor are the TVERC data available as per the licence agreement.