Measure execution speed inside of Shiny - r

Question

I am working on developing a Shiny app.
I am interested in clocking the time it takes to execute certain chunks of code (such as a ggplot, etc).
For some reason it appears that using the usual clocking methods don't work within reactive calls, for example:
output$R1_C1 <- renderPlot({
beginning <- Sys.time()
<lots of code here>
end <- Sys.time()
print(end - beginning)
R complains and gives me
Error in (structure(function (input, output) :
object 'beginning' not found
Has anyone found a successful way to time execution speed inside of reactive calls in Shiny?

Answer 1

This works on my system:
library(shiny)
runApp(list(
ui = bootstrapPage(
numericInput('n', 'Number of obs', 100),
plotOutput('plot')
),
server = function(input, output) {
output$plot <- renderPlot({
beginning <- Sys.time()
h <- hist(runif(input$n))
end <- Sys.time()
print(end - beginning)
h
})
}
))

Answer 2

In a similar way, you can use the tictoc package as shown below. The reason I prefer this over the base R solution is because you can easily use multiple tic/toc tags to measure sub-routines within either a reactive or a rendering expression.
library(shiny)
library(tictoc)
runApp(list(
ui = bootstrapPage(
numericInput('n', 'Number of obs', 100),
plotOutput('plot')
),
server = function(input, output) {
output$plot <- renderPlot({
tic("execution time - Histogram")
hist(runif(input$n))
toc()
tic("execution time - Print")
print("this is a second task within 'renderPlot' ")
toc()
})
}
))

Answer 3

The profvis package might be useful. Example:
library(shiny)
library(profvis)
profvis({
sApp <- shinyApp(
ui = fluidPage(
numericInput('n', 'Number of obs', 100, min = 1, max = 200),
plotOutput('plot')
),
server = function(input, output) {
dfTable <- reactive({
as.data.frame(matrix(rnorm(10 * input$n, mean = 5), ncol = input$n))
})
vMeans <- reactive({
apply(dfTable(), 2, mean)
})
output$plot <- renderPlot({
hist(vMeans())
})
}
)
runApp(sApp)
})