R routines for unstructured interpolation. A few methods are implemented which work relatively efficiently on semi-large datasets (e.g. millions of points).
The code relies heavily on other packages (FNN + geometry), and its purpose is mainly to provide a convenient interface to some efficient interpolation methods (and save us hacking to use them elsewhere!), rather than on developing novel interpolation methods.
Note that many other R packages have routines for unstructured interpolation, and only a few methods are supported here.
Make sure you have the 'devtools', 'FNN', and 'geometry' package installed in R:
install.packages(c('devtools','FNN','geometry'))
Then you can install direct from github
install_github('GeoscienceAustralia/unstructInterp')
To get started, look at the examples (which also function as unit tests) and help:
library(unstructInterp)
example(nearest_neighbour_interpolation)
?nearest_neighbour_interpolation
example(triangular_interpolation)
?triangular_interpolation
This code is sporadically developed based on the needs of the developers and colleagues. The open source release is mainly to assist with our own management of the code, as well as to allow others to use it. Bugfixes/suggestions are nonetheless welcome.
If you'd like to contribute any routines or fix bugs, consider forking the code, inserting your changes and sending the maintainer a pull request. Otherwise bug reports / suggestions can be sent to the maintainer Gareth Davies, gareth.davies.ga.code@gmail.com
Note that the documentation in /man is auto-generated from within-code comments using the roxygen2 package.
This code is licensed under a BAD 3-clause license. See the license deed for details
Gareth Davies
Lead Devloper
gareth.davies@ga.gov.au