Search code examples
rggplot2plotlyr-plotlyggplotly

sunburst chart for multi columns in R using shiny

Consider following data

 > temp
   year  v13 per_v10 per_v11 per_v12
 1 1397 2506   20.11   65.13   14.76
 2 1398 1759   16.66   79.99    3.35
 3 1399 1754   19.44   75.43    5.13

 temp<-structure(list(year = c(1397, 1398, 1399), v13 = c(2506, 1759, 
 1754), per_v10 = c(20.11, 16.66, 19.44), per_v11 = c(65.13, 79.99, 
 75.43), per_v12 = c(14.76, 3.35, 5.13)), row.names = c(NA, -3L
 ), class = "data.frame")

I want to make a pie plot for 4 variables (per_v10 ,per_v11 ,per_v12 ,per_v13) in three years(1397,1398,1399). It should be something like this:

enter image description here

with Onclick in each year(1397:1399) it shows the percent only for that year. I saw this , but it is too complicated for me to understand.

Solution 
  •  temp<-structure(list(year = c(1397, 1398, 1399), v13 = c(2506, 1759, 
 1754), per_v10 = c(20.11, 16.66, 19.44), per_v11 = c(65.13, 79.99, 
 75.43), per_v12 = c(14.76, 3.35, 5.13)), row.names = c(NA, -3L
 ), class = "data.frame")
 
# create a df
(df <- data.frame(
     ids = c("total", paste0("total - ", temp$year), paste0("total - ", rep(temp$year, each = 3), " - ", rep(c('per_v10', 'per_v11', 'per_v12'), 3))),
     labels = c("Total", as.character(temp$year), rep(c('per_v10', 'per_v11', 'per_v12'), 3)),
     parents = c("", "total", "total", "total", rep(paste0("total - ", temp$year), each =3)),
     values = c(rep(0, 4), c(t(temp[, 3:5])))
 ))

library(plotly)
library(shiny)
library(shiny)
 
 ui <- fluidPage(
   plotlyOutput("myplot"),
   textOutput("percent")
 )
 
 server <- function(input, output, session) {
   output$myplot <- renderPlotly({
       plot_ly(
           data = df,
           ids = ~ids,
           labels = ~labels,
           parents = ~parents,
           type = 'sunburst',
           values = ~values,
           source = "myplot_source"
       )
   })
   output$percent <- renderPrint({
        clicked <- event_data(event = "plotly_click", source = "myplot_source", priority = "event")$pointNumber + 1
        req(clicked)
        if (clicked < 5) "You are not clicking on the children"
        else paste0("Clicked percentage is ", df$values[clicked])    
   })
 }
 
shinyApp(ui, server)

    I plot the most outside layer width by the percentage value you provided. If you want each bin to be evenly distributed, remove the line values = ~values,

    Not sure what do you mean by Onclick in each year. So I made it like the link you provided. When you clicked on the bins, a text bar below will be displayed the percentage of the clicked bins.

    enter image description here

    Plotting sunburst in R is annoying, you need to reformat the dataframe. I suggest you try out the Python APIs.