Clustered Bar chart in R shiny

[kuttan98]

Hi,
I have a data for monitoring teachers. There are master trainers and under them we have teachers. In the shiny app I need clustered bar chart of score of teachers week-wise. In the X-axis, I will have the name of teachers and y-axis the average score. I want the graph to react to the trainer and week. I select the trainer first and then the week, the graphs of that week will appear for all the teachers. This is how how I want it to appear.

library(tidyverse)
library(shiny)
library(shinydashboard)
#> 
#> Attaching package: 'shinydashboard'
#> The following object is masked from 'package:graphics':
#> 
#>     box
library(janitor)
#> 
#> Attaching package: 'janitor'
#> The following objects are masked from 'package:stats':
#> 
#>     chisq.test, fisher.test
library(plotly)
#> 
#> Attaching package: 'plotly'
#> The following object is masked from 'package:ggplot2':
#> 
#>     last_plot
#> The following object is masked from 'package:stats':
#> 
#>     filter
#> The following object is masked from 'package:graphics':
#> 
#>     layout
data<-tibble::tribble(
      ~trainer, ~week_no,    ~teacher, ~score,
  "Madhumathi",       1L,    "Gandhi",    76L,
       "Nagma",       1L,     "Nehru",    83L,
        "Anil",       1L,     "Patel",    86L,
  "Madhumathi",       1L,    "Rajaji",    57L,
       "Nagma",       1L,     "Obama",    73L,
        "Anil",       1L, "Roosevelt",    67L,
  "Madhumathi",       2L,    "Gandhi",    73L,
       "Nagma",       2L,     "Nehru",    86L,
        "Anil",       2L,     "Patel",    72L,
  "Madhumathi",       2L,    "Rajaji",    65L,
       "Nagma",       2L,     "Obama",    70L,
        "Anil",       2L, "Roosevelt",    84L,
  "Madhumathi",       3L,    "Gandhi",    75L,
       "Nagma",       3L,     "Nehru",    90L,
        "Anil",       3L,     "Patel",    70L,
  "Madhumathi",       3L,    "Rajaji",    75L,
       "Nagma",       3L,     "Obama",    70L,
        "Anil",       3L, "Roosevelt",    63L
  )

ui<-dashboardPage(
  skin = "red",
  dashboardHeader(title = "EarlySpark MIS Dashboard 2022-23",titleWidth = 550),
  dashboardSidebar("Choose your inputs here",
                   selectInput("master","Select the Master Trainer",choices=unique(data$trainer),multiple=T),
                   selectInput("week","Select the Week number", choices = NULL,multiple = T)),
  dashboardBody(
    tabsetPanel(
      tabPanel(title = "First Tab",
               plotlyOutput("plot1",height = 500,width = 1500))
      
    )
  )
)


server<-function(input,output,session){
  observe({
    req(input$master)
    
    
    x<-data %>% 
      filter(trainer %in% input$trainer) %>% 
      select(week_no)
    updateSelectInput(session,"week","Select the Week number",choices = c("All",x))
  })
  
  instr_attend<-reactive({  
    req(input$master)
    req(input$week)
    
    data %>% 
      group_by(trainer,week_no,teacher,score) %>% 
      summarise(avg_score=mean(score)) %>% 
      filter(trainer %in% input$master | "All"  %in% input$master,
             week_no %in% input$week | "All" %in% input$week) 
  })
  
  output$plot1<-renderPlotly({
    req(instr_attend())
    
    student_att1<-ggplot(instr_attend(),aes(teacher,score,fill=week_no))+
      geom_bar(stat="identity",width = 0.5,position = "dodge")+
      coord_flip()+
      theme_minimal()
    
    ggplotly(student_att1)
  })
  
}

shinyApp(ui,server)
#> PhantomJS not found. You can install it with webshot::install_phantomjs(). If it is installed, please make sure the phantomjs executable can be found via the PATH variable.
Shiny applications not supported in static R Markdown documents
Created on 2022-09-20 by the reprex package (v2.0.1)

[dvetsch75]

It's a bit unclear to me what exactly you are looking for... I ran your app and I do get a (horizontal) barchart.

I think if it were me, I would probably do the following:

• Get rid of the coord_flip in your ggplot. That should correct the axes to the teacher being on the x and the score on the y.
• Turn your scores into categorical variables instead of continuous ones. The categorical makes position_dodge a little bit more sensible to ggplot. In the code snippet below, I've done that using as.factor.
• Use geom_col instead of geom_bar. While you can override the stat for geom_bar, I think it would be a lot easier to introduce bugs into your code by specifying one type of plot and changing the way it aggregates data when there is another type of plot that always aggregates data the way you want. From the docs for geom_bar:

If you want the heights of the bars to represent values in the data, use geom_col() instead.

So with that being said, here is probably how I would change your plot:

(instr_attend() %>% 
            ggplot(aes(x = teacher, y = avg_score)) +
            geom_col(aes(fill = as.factor(week_no)), position = 'dodge') + 
            scale_fill_discrete(name = 'Week No') +
            theme_minimal()) %>% 
            ggplotly()

Which yields this:

[kuttan98]

Thank you very much. The app was working as intended. Sorry for that. But the additional inputs you have given is really helpful.

Regards,
NP