In R, there is a wide range of libaries used for linear regression. This often makes R scripts difficult to read given repeating code blocks. LinReg is a tool box I am continously developing to allow for easy analysis of various datasets.
The backbone of LinReg contains the most common regression libaries (lmer, glmer, glmmTMB,nlme, etc) as well as helper packages (emmeans, mass, car, statmod, multcomp, ggplot,effsize, etc) in a single spot.Statistical pipelines should be a common set of questions that make up a common protocol. These question might include:
- Is the data normally distributed?
- Does this dataset violate any assumptions for linear regression?
- Should my model contain random effects?
- How should we handle missing data?
- How well does the model fit the data?
- Should I use a hold-out dataset to validate model performance?
LinReg attempts to answer these questions sequentially and allow users to easily develop models from multiple packages utalizing a single library.