shinybrms

Description

The R package shinybrms provides a graphical user interface (GUI) for fitting Bayesian regression models using the R package brms which in turn relies on Stan. The shinybrms GUI is a shiny app.

To get an impression of the shinybrms app, see the Get Started page.

Launching the shinybrms app

The following two sections describe two ways for launching the shinybrms app, either with or without the installation of shinybrms. The former is recommended as it offers all advantages that R packages have (e.g., offline usage). For both ways, the following steps need to be performed first:

1. Install R (see the R homepage).
2. Install the R package rstan (see the RStan Getting Started GitHub page for instructions; make sure to use install.packages("rstan", [...], dependencies = TRUE) with [...] as advised there).
3. If the cmdstanr backend should be used (or if it is needed because the rstan backend does not work as expected), then cmdstanr as well as CmdStan need to be installed by following the instructions on the cmdstanr homepage. In general, however, the rstan backend should be sufficient. In the context of shinybrms, the major advantage of the cmdstanr backend compared to the rstan backend is that in general, it provides a speed improvement for the Stan run.

With installation of shinybrms

1. Use one of the following approaches to install the R package shinybrms either from CRAN or from GitHub. The GitHub version might be more recent than the CRAN version, but the CRAN version might be more stable. Another decision is whether the example datasets from the R packages lme4, MASS, and rstanarm should be available.

   • If the example datasets from the R packages lme4, MASS, and rstanarm should be available, then the R code for installing shinybrms from CRAN or GitHub (respectively) is as follows:

     • To install shinybrms from CRAN:

       install.packages("shinybrms", dependencies = TRUE)

     • To install shinybrms from GitHub:

       if (!requireNamespace("devtools", quietly = TRUE)) {
         install.packages("devtools")
       }
       devtools::install_github("fweber144/shinybrms", dependencies = TRUE)

   • If the example datasets from the R packages lme4, MASS, and rstanarm should not be available, then the R code for installing shinybrms from CRAN or GitHub (respectively) is as follows:

     • To install shinybrms from CRAN:

       install.packages("shinybrms")

     • To install shinybrms from GitHub:

       if (!requireNamespace("devtools", quietly = TRUE)) {
         install.packages("devtools")
       }
       devtools::install_github("fweber144/shinybrms")

2. Launch the shinybrms app by either running the following R code:

   library(shinybrms)
   launch_shinybrms()

   or this R code which ensures that the app opens up in the default web browser (helpful, e.g., if RStudio is used):

   library(shinybrms)
   launch_shinybrms(launch.browser = TRUE)

Without installation of shinybrms

1. Install the R package brms. The following R code may be used for this:

   install.packages("brms")

2. If the example datasets from the R packages lme4, MASS, and rstanarm should be available, these packages need to be installed. The following R code may be used for this:

   install.packages(c("lme4", "MASS", "rstanarm"))

3. Launch the shinybrms app directly from GitHub by either running the following R code:

   shiny::runGitHub("fweber144/shinybrms",
                    subdir = "inst/shinybrms_app")

   or this R code which ensures that the app opens up in the default web browser (helpful, e.g., if RStudio is used):

   shiny::runGitHub("fweber144/shinybrms",
                    subdir = "inst/shinybrms_app",
                    launch.browser = TRUE)