useful tools

Pipelining tools

To automate my analyses in R, I became fan of targets (formerly: drake), which is similar to Make. After some onboarding headaches, I can’t imagine working without it for new analysis projects.

Use Julia and R combined

Julia is extremely fast in fitting Mixed models (and probably also other things). For large models, I recommend checking out using Julias MixedModels.jl, which uses the same notations than R’s lme4/lmerTest package. To seamlessly integrate Julia in my R workflow, I use the JuliaCall package in R. To be able to convert Julia models to R model objects, the Julia JellyMe4 package does a wonderful job. All R functions applied to an lmer object can likewise be applied to such a converted Julia model. Some hustle can happen, when the model to be converted has special properties, such as zerocorr / ||, but the Julia community is very helpful.

Quarto

Use quarto to make a website on GitHub (youtube tutorial series), and automatically render website on push.