Shiny Exercise 1

These exercises are about building basic apps, app layouts, and themes.

Exercises for Shiny session 1

Make an app that reads in the table in the ‘data’ folder of the course materials that has a file name of ‘msk_2024_clinical_Pancreatic_rev.csv’ and displays that table on the page.

library(ggplot2)
library(dplyr)
library(shiny)
library(bslib)
library(DT)
library(survival)


data_in <- rio::import("data/msk_2024_clinical_Pancreatic_rev.csv")

ui <- page_fluid(
  card(
    card_header("Clinical Data"),
    DT::dataTableOutput("data_in")
    )
)

server <- function(input, output){
  
  output$data_in <- renderDataTable({
    datatable(data_in)
  })
  
  
}

shinyApp(ui, server)

Add some figures to the page, including a plot that:

displays a histogram showing the distribution of the ages of patients
displays three pie charts showing the break down of the following columns: Sex, Sample_Type, and Smoking_History_NLP
displays one Kaplan Meier survival curve showing any difference in survival based on the Smoking_History_NLP column
- HINT: see sample code below for how to make a histogram, pie chart, and survival curve in the code chunk below.
Put the table and plots in separate cards and make the card that has the table collapsable.

# SAMPLE CODE FOR PLOTS:

data_in <- rio::import("data/msk_2024_clinical_Pancreatic_rev.csv")

# Sample code for pie chart:
library(ggplot2)
library(dplyr)

data_in %>%
  group_by(Sex) %>%
  summarize(Num = n()) %>%
  ggplot(aes(x = 1, y = Num, fill = Sex)) +
  geom_bar(stat="identity", color="white") +
  coord_polar("y") + theme_void()

# Sample code for histogram:
ggplot(data_in, aes(x = Current_Age)) +
      geom_histogram(bins=10) + 
      theme_bw()

# Sample code for survival analysis:
library(survival)

surv_obj <- Surv(time = data_in$Overall_Survival_Months, event = data_in$Survival_code)
fit <- survfit(surv_obj ~ Smoking_History_NLP, data = data_in)
surv_df <- data.frame(
  time = fit$time,
  surv = fit$surv,
  group = rep(names(fit$strata), fit$strata)
)

ggplot(surv_df, aes(x = time, y = surv, color = group)) +
  geom_step(linewidth = 1) +
  labs(x = "Time (months)", y = "Survival Probability", color = "Group") +
  theme_minimal()

library(ggplot2)
library(dplyr)
library(shiny)
library(bslib)
library(DT)
library(survival)

data_in <- rio::import("data/msk_2024_clinical_Pancreatic_rev.csv")

surv_obj <- Surv(time = data_in$Overall_Survival_Months, event = data_in$Survival_code)
fit_smokes <- survfit(surv_obj ~ Smoking_History_NLP, data = data_in)
surv_df <- data.frame(
  time = fit_smokes$time,
  surv = fit_smokes$surv,
  group = rep(names(fit_smokes$strata), fit_smokes$strata)
)

ui <- page_fluid(
  card(
    card_header("Clinical Data"),
    DT::dataTableOutput("data_in")),
  card(card_header("Age distribution"),
       plotOutput("age_histogram")),
  card(card_header("Smoking status effect on survival"),
       plotOutput("smokes_kaplan")),
  card(card_header("Sex"),
       plotOutput("sex_pie_chart")),
  card(card_header("Smoking status"),
       plotOutput("smokes_pie_chart")),
  card(card_header("Sample Type"),
       plotOutput("sampleType_pie_chart")),
  card(card_header("Cancer Type"),
       plotOutput("cancerType_pie_chart"))
)

server <- function(input, output){
  
  output$data_in <- renderDataTable({
    datatable(data_in)
  })
  
  output$age_histogram <- renderPlot({
    ggplot(data_in, aes(x = Current_Age)) +
      geom_histogram(bins=10) + 
      theme_bw()
  })
  
  
  output$sex_pie_chart <- renderPlot({
    data_in %>%
      group_by(Sex) %>%
      summarize(Num = n()) %>%
      ggplot(aes(x = 1, y = Num, fill = Sex)) +
      geom_bar(stat="identity", color="white") +
      coord_polar("y") + theme_void()
  })
  
  output$sampleType_pie_chart <- renderPlot({
    data_in %>%
      group_by(Sample_Type) %>%
      summarize(Num = n()) %>%
      ggplot(aes(x = 1, y = Num, fill = Sample_Type)) +
      geom_bar(stat="identity", color="white") +
      coord_polar("y") + theme_void()
  })
  
  output$smokes_pie_chart <- renderPlot({
    data_in %>%
      group_by(Smoking_History_NLP) %>%
      summarize(Num = n()) %>%
      ggplot(aes(x = 1, y = Num, fill = Smoking_History_NLP)) +
      geom_bar(stat="identity", color="white") +
      coord_polar("y") + theme_void()
  })
  
  output$cancerType_pie_chart <- renderPlot({
    data_in %>%
      group_by(Cancer_Type_Detailed) %>%
      summarize(Num = n()) %>%
      ggplot(aes(x = 1, y = Num, fill = Cancer_Type_Detailed)) +
      geom_bar(stat="identity", color="white") +
      coord_polar("y") + theme_void()
  })
  output$smokes_kaplan <- renderPlot({
    ggplot(surv_df, aes(x = time, y = surv, color = group)) +
      geom_step(linewidth = 1) +
      labs(x = "Time (months)", y = "Survival Probability", color = "Group") +
      theme_minimal()
  })
  
}

shinyApp(ui, server)

Make the app have multiple pages, with the first page containing all of the plots and the second page containing the table.

For the page with the plots, try various layouts with at least one row having multiple columns and pick your favorite layout.
add a menu panel on the right side of the nav bar with a link to the page where these data came from, https://www.cbioportal.org/
since the table and plots are now on separate pages, add a value_box to the page with the plots to let the user quickly know the number of patients in this study.
Try a couple of different themes and pick one. Remember that you often need to restart your R session to clear the cache and reset the theme.

library(ggplot2)
library(dplyr)
library(shiny)
library(bslib)
library(DT)
library(survival)


data_in <- rio::import("data/msk_2024_clinical_Pancreatic_rev.csv")

surv_obj <- Surv(time = data_in$Overall_Survival_Months, event = data_in$Survival_code)
fit_smokes <- survfit(surv_obj ~ Smoking_History_NLP, data = data_in)
surv_df <- data.frame(
  time = fit_smokes$time,
  surv = fit_smokes$surv,
  group = rep(names(fit_smokes$strata), fit_smokes$strata)
)


ui <- page_navbar(
  title = "Cancer data",
  theme = bs_theme(version = 5, bootswatch = "pulse"),
  nav_panel(
    title = "Data visualization",
    
    layout_columns(
      layout_columns(
        value_box("Number of patients", 
                  value = nrow(data_in),
                  theme = "text-purple"),
        card(card_header("Age distribution"),
             plotOutput("age_histogram")),
        col_widths = c(3,9), 
        row_heights = list("300px")
      ),
      layout_columns(
        card(card_header("Sex"),
             plotOutput("sex_pie_chart")),
        card(card_header("Smoking status"),
             plotOutput("smokes_pie_chart")),
        card(card_header("Sample Type"),
             plotOutput("sampleType_pie_chart")),
        card(card_header("Cancer Type"),
             plotOutput("cancerType_pie_chart")),
        col_widths = c(6, 6, 6, 6),
        row_heights = c("400px", "400px")
      ),
      layout_columns(
        card(card_header("Survival by group"),
             plotOutput("smokes_kaplan")),
        col_widths = c(12), 
        row_heights = list("500px")
      ),
      col_widths = c(12,12, 12), fill=F
    )
    
  ),
  nav_panel(
    title = "Clinical data table",
    card(
      card_header("Clinical Data"),
      DT::dataTableOutput("data_in")
    )
  ),
  nav_spacer(),
  nav_menu(
    title = "Links",
    align = "right",
    nav_item(
      tags$a(
        shiny::icon("chart-simple"), "RU BRC - Learn more!",
        href = "https://www.cbioportal.org/",
        target = "_blank"
      )
    )
  )
)

server <- function(input, output){
  
  output$data_in <- renderDataTable({
    datatable(data_in)
  })
  
  output$age_histogram <- renderPlot({
    ggplot(data_in, aes(x = Current_Age)) +
      geom_histogram(bins=10) + 
      theme_bw()
  }) 
  
  
  output$sex_pie_chart <- renderPlot({
    data_in %>%
      group_by(Sex) %>%
      summarize(Num = n()) %>%
      ggplot(aes(x = 1, y = Num, fill = Sex)) +
      geom_bar(stat="identity", color="white") +
      coord_polar("y") + theme_void()
  })
  
  output$sampleType_pie_chart <- renderPlot({
    data_in %>%
      group_by(Sample_Type) %>%
      summarize(Num = n()) %>%
      ggplot(aes(x = 1, y = Num, fill = Sample_Type)) +
      geom_bar(stat="identity", color="white") +
      coord_polar("y") + theme_void()
  })
  
  output$smokes_pie_chart <- renderPlot({
    data_in %>%
      group_by(Smoking_History_NLP) %>%
      summarize(Num = n()) %>%
      ggplot(aes(x = 1, y = Num, fill = Smoking_History_NLP)) +
      geom_bar(stat="identity", color="white") +
      coord_polar("y") + theme_void()
  })
  
  output$cancerType_pie_chart <- renderPlot({
    data_in %>%
      group_by(Cancer_Type_Detailed) %>%
      summarize(Num = n()) %>%
      ggplot(aes(x = 1, y = Num, fill = Cancer_Type_Detailed)) +
      geom_bar(stat="identity", color="white") +
      coord_polar("y") + theme_void()
  })
  
  output$smokes_kaplan <- renderPlot({
    ggplot(surv_df, aes(x = time, y = surv, color = group)) +
      geom_step(linewidth = 1) +
      labs(x = "Time (months)", y = "Survival Probability", color = "Group") +
      theme_minimal()
  })
  
  
}

shinyApp(ui, server)

Shiny Exercise 1

Rockefeller University, Bioinformatics Resource Centre

https://rockefelleruniversity.github.io/RU_shiny/