The flocker R package enables users to fit flexible occupancy models using
the extended lme4 formula syntax provided by R package brms. Also
included are functions to format data for occupancy modeling and functions
for model post-processing, including posterior predictions, the posterior
distribution of the latent true occupancy state, and model comparison.
flocker is under active development: development priorities include better
direct functionality for posterior predictive checking and implementation of
dynamic (multiseason) occupancy models and N-mixture models.
Compared to existing R packages for fitting occupancy models, flocker is
substantially more flexible in the variety of single-season models that can
be fit, and contains advanced functionality for model comparison and
checking.
To get started, check out our tutorial vignette, available here. For installation instructions, see below.
Install flocker in R with
# install.packages("remotes")
remotes::install_github("jsocolar/flocker")
flocker requires a working version of either Rstan or cmdstan, which are
interfaces to Stan, a state-of-the-art the
probabilistic programming language. We recommend using cmdstan if possible.
To do so, first install R package cmdstanr with
install.packages("cmdstanr", repos = c("https://mc-stan.org/r-packages/", getOption("repos")))
You must additionally install cmdstan itself. We strongly recommend using
cmdstanr to manage the cmdstan installation:
cmdstanr::install_cmdstan()
If using Rstan, we strongly recommend installing the development version rather than the CRAN version. In a fresh R session, do
# Uncomment the next line if you have previously installed rstan
# remove.packages(c("StanHeaders", "rstan"))
install.packages("StanHeaders", repos = c("https://mc-stan.org/r-packages/", getOption("repos")))
install.packages("rstan", repos = c("https://mc-stan.org/r-packages/", getOption("repos")))
Both Rstan and cmdstan require a working C++ toolchain, which has posed occasional complications for Stan users. See here for Rstan-focused advice on installing the toolchain, and see here for cmdstan-focused advice. If you encounter toolchain issues that you are unable to troubleshoot, ask for help at https://discourse.mc-stan.org/.
To request a feature or report a bug (much appreciated!) please
open an issue at the GitHub repository.
To contribute to flocker (very much appreciated!) have a look at existing open
issues, or open a new issue to discuss.
Please cite flocker as:
- Socolar, J.B. & Mills, S.C. (2021). "flocker: flexible occupancy estimation in R." R package version XXX, <URL: https://github.com/jsocolar/flocker/>.
Additionally, please consider citing Stan and R package brms.