Shinyscholar

A template for creating reproducible Shiny applications in R

Simon Smart and Tim Lucas, School of Medical Sciences, University of Leicester

9th September 2025

Shinyscholar helps create reproducible analytical apps in R

• Creates an empty application with a regular structure
• Analyses can be reproduced outside the app

Shinyscholar helps create reproducible analytical apps in R

• Creates an empty application with a regular structure
• Analyses can be reproduced outside the app
• Make it easier to follow software development best practices
• Developers can concentrate on creating functionality

Testing, modularised, maintainable

R has cutting-edge statistical methods but barriers can restrict uptake

• Know how to use R

R has cutting-edge statistical methods but barriers can restrict uptake

• Know how to use R
• Able to comprehend documentation

R has cutting-edge statistical methods but barriers can restrict uptake

• Know how to use R
• Able to comprehend documentation
• Access to specific data

R has cutting-edge statistical methods but barriers can restrict uptake

• Know how to use R
• Able to comprehend documentation
• Access to specific data
• Format data in specific ways

Especially targeting technical users, but who might not code. Conservationists, analysts, medical practitioners

Shiny apps enable anybody to access the power of R

ui <- fluidPage(
  numericInput("number", "Enter a number", value = 5),
  textOutput("answer")
)

server <- function(input, output) {
  output$answer <- renderText(input$number * 10)
}

shinyApp(ui, server)

Shiny apps are becoming increasingly popular in academia

• The low barrier to entry makes Shiny popular
• Substantially increased apps as a method of dissemination

• Easier to develop than using other web frameworks
• Existing code can be adapted

Current apps do not follow best practices

Wallace bucked the trend

Kass et al. (2022) DOI: 10.1111/ecog.06547 install.packages("wallace") (not online)

• Error messages

• Logging

• Save and load

• Reproducible

• Cites packages used

• Take the lessons learnt from them and make available to all

• We were going to reuse Wallace as the basis for a new app, so it seemed to make sense to make it available to others

• Lead onto modules here - i.e. occ_queryDb

modules contains information about each module

modules <- data.frame(
"component" = "transform",
"long_component" = "Transform numbers",
"module" = "multiply",
"long_module" = "Multiply a number",
"map" = FALSE,
"result" = TRUE,
"rmd" = TRUE,
"save" = TRUE,
"download" = FALSE,
"async" = FALSE)

create_template() creates the skeleton app

shinyscholar::create_template(
  path = file.path("~", "Documents"), 
  name = "demo", 
  author = "Simon Smart",
  include_map = FALSE, 
  include_table = FALSE, 
  include_code = FALSE, 
  common_objects = "result", 
  modules = modules, 
  install = TRUE)

demo::run_demo()

create_template() creates the skeleton app

Reproducibility relies on each module calling one function

Create the function:

transform_multiply <- function(number) number * 10

Reproducibility relies on each module calling one function

Create the function:

transform_multiply <- function(number) number * 10

Use it in the module and store the input:

common$result <- transform_multiply(input$number)
common$meta$transform_multiply$number <- input$number

Reproducibility relies on each module calling one function

Create the function:

transform_multiply <- function(number) number * 10

Use it in the module and store the input:

common$result <- transform_multiply(input$number)
common$meta$transform_multiply$number <- input$number

In the Rmarkdown:

result <- transform_multiply({{transform_multiply_number}})

Reproducibility relies on each module calling one function

Create the function:

transform_multiply <- function(number) number * 10

Use it in the module and store the input:

common$result <- transform_multiply(input$number)
common$meta$transform_multiply$number <- input$number

In the Rmarkdown:

result <- transform_multiply({{transform_multiply_number}})

metadata() takes care of a lot of the boring code

Disagapp for disaggregation regression

MetaInsight for network metaanalysis

Learn more

• install.packages("shinyscholar")
• Example app at https://simonsmart.shinyapps.io/shinyscholar/ install.packages("shinyscholar", dependencies = TRUE)
• README guide to development at https://simon-smart88.github.io/shinyscholar/
• Workshop for Ukraine

Acknowledgments

• Wellcome for funding
• Wallace developers especially Rob Anderson
• Tim Lucas