R Shiny: Modify output values in a data frame - r

Question

The following simple shiny app displays a word and its sentiment as stored in the R data frame named sent.
library(shiny)
sent <- data.frame(word=c('happy', 'sad', 'joy', 'upset'),
sentiment=c('positive', 'negative', 'positive', 'negative'),
stringsAsFactors = FALSE)
ui <- fluidPage(
numericInput(inputId = 'num', label='', value=1, min=1, max=nrow(sent)),
br(),
h4("Word:"),
textOutput('word'),
br(),
h4("Sentiment:"),
textOutput('sentiment')
)
server <- function(input, output){
output$word <- renderText({ sent$word[input$num] })
output$sentiment <- renderText({ sent$sentiment[input$num] })
}
shinyApp(ui=ui, server=server)
I would like to modify this in 2 ways:
(1) I would like the user to be able to scroll through the words in the column sent$word, rather than using numericInput()
(2) More importantly, I would like the user to be able to modify the sentiment value associated with each word. Ideally, this would be a drop down menu (with 'positive' and 'negative' as options), which would display the current sentiment value stored in sent for that word, but which could be changed by the user and overridden in the dataframe.
Any suggestions?

Answer 1

This should do the trick
library(shiny)
sent <- data.frame(word=c('happy', 'sad', 'joy', 'upset'),
sentiment=c('positive', 'negative', 'positive', 'negative'),
stringsAsFactors = FALSE)
sent2 <- reactiveVal(sent)
i <- 1
i2 <- reactiveVal(i)
ui <- fluidPage(
uiOutput("wordSelect"),
br(),
h4("Word:"),
textOutput('word'),
br(),
h4("Sentiment:"),
textOutput('sentiment'),
br(),
uiOutput("change"),
actionButton("go","Change")
)
server <- function(input, output){
output$wordSelect <- renderUI({
selectizeInput(inputId = 'wrd', label='select word', choices=sent$word, selected=sent$word[i2()])
})
output$word <- renderText({ input$wrd })
output$sentiment <- renderText({ sent$sentiment[which(sent2()$word==input$wrd)] })
observeEvent(input$go, {
out <- sent
out$sentiment[which(sent$word==input$wrd)] <- input$newLabel
sent <<- out
sent2(out)
i <<- which(sent$word==input$wrd)+1
if(i > length(sent$word)) {
i <<- i - 1
}
i2(i)
})
output$change <- renderUI({
radioButtons("newLabel", label="Change value", choices=c('positive','negative'), sent$sentiment[which(sent2()$word==input$wrd)])
})
}
shinyApp(ui=ui, server=server)
The adjusted output is first stored in a reactiveVal named sent2(). This is required for you see the adjure values while running the Shiny App.
A selectizeInput() is used to scroll through the words (Q1).
radioButtons() are used to select positive and negative values. The default value is whatever value is currently applied to the corresponding word.
An actionButton() is used to make the change when wanted.
UPDATE: I added sent <<- out so that your sent dataframe actually gets updated. Be aware that this will overwrite the values you had stored in sent before.
UPDATE: Each time the action button is clicked, the index of the currently selected word is determined using which(). Then it is incremented and stored in i and i2(). The new index is used to determine the default value of selectizeInput(). This way, when no manual selection of words is done, you will scroll through all options. When a word is selected manually, you will continue incrementing from that word onwards. When the last word is reached, the value does not increment further