Software
We develop our research software in the open. Code and data are public goods of science: publishing them makes our results transparent and reproducible, and lets others build on them rather than start from scratch.
Below are the R packages we develop or contribute to. Code and data that accompany individual papers are on the group software page, in the Theoretical Ecology GitHub organisation, on my GitHub account, and on the pages of individual group members.
Main packages
DHARMa
Simulation-based residual diagnostics for generalized linear mixed models and other hierarchical regression models.
BayesianTools
MCMC and SMC samplers and diagnostics for Bayesian statistics, with a focus on calibrating complex process-based models.
Further packages from the group
sjSDM
Fast joint species distribution models for large community datasets, estimating the full species covariance matrix.
cito
Deep neural networks in R with ordinary formula syntax, plus the explainable-AI tools to interpret them.
Contributions to other projects
rLPJGUESS
An R interface to the LPJ-GUESS vegetation model, with parallel runs on workstations and HPC clusters.
ProfoundData
An R interface to the PROFOUND database of European forest data for evaluating vegetation models.
r3PG
A fast R interface to the 3-PG forest growth model, for pure and mixed stands.
NUCOMBog
An R package of the NUCOM-Bog model of vegetation, carbon, nitrogen and water dynamics in peatlands.
gen3sis
A general engine for spatially explicit eco-evolutionary simulations of biodiversity.