The GARP modelling system: problems and solutions to automated spatial prediction

TLDR

The paper addresses challenges in reliably analyzing arbitrary spatial datasets, noting that statistical issues in such data are broadly applicable across domains. The authors aim to describe components of the GARP Modelling System (GMS) for automating predictive spatial modelling of species distributions. GMS comprises a generic spatial modelling method that filters potential errors and is integrated with existing database, visualization tools, and web browsers to facilitate development. The GMS has proven successful in delivering user‑friendly spatial data analysis over the Internet.

Abstract

This paper is concerned with the problems and solutions to reliable analysis of arbitrary datasets. Our approach is to describe components of a system called the GARP Modelling System (GMS) which we have developed for automating predictive spatial modelling of the distribution of species of plants and animals. The essence of the system is an underlying generic spatial modelling method which filters out potential sources of errors. The approach is generally applicable however, as the statistical problems arising in arbitrary spatial data analysis potentially apply to any domain. For ease of development, GMS is integrated with the facilities of existing database and visualization tools, and Internet browsers. The GMS is an example of a class of application which has been very successful for providing spatial data analysis in a simple to use way via the Internet.