app.R

# Packages ----------------------------------------------------
library(shiny)
library(shinydashboard)

library(tidyverse)
library(janitor)
library(grid)

library(udpipe)
library(wordcloud)
library(bsicons)
library(htmltools)

library(fontawesome)
library(DT)

library(googlesheets4)
gs4_deauth()

# Time interval
time_interval <- 604800000

# Everyone, Leadership, new to data science, software developers
cbPalette <- c("#E69F02", "#56B4E9", "#009E73", "#008080")

#Leadership, new to data, software developers
cbLEPalette <-c("#56B4E9", "#009E73", "#008080", "#440154")

xlabel_view <- c(rep(c("black", "transparent", "transparent", "transparent"), 41), "black", "transparent") #166 rows
#cc <- rev(c("#fde725", "#addc30", "#5ec962", "#28ae80", "#21918c", "#2c728e", "#3b528b", "#472d7b", "#440154"))
viridis_cc <- c("#440154", "#2c728e", "#fde725", "#28ae80", "#addc30")

# Wordcloud
ud_model <- udpipe::udpipe_load_model("wordcloud-model.udpipe")

# Functions

coacross <- function(...) {
  coalesce(!!!across(...))
}

get_to_bind <- function(inputdf, prepost, workshopOI){
  nrankOI <- nrow(inputdf %>%
                    filter(pre_post == prepost & workshop == workshopOI) %>% drop_na())
  
  return(inputdf %>%
           filter(pre_post == prepost & workshop == workshopOI) %>%
           select(value) %>% `colnames<-`(c(paste(workshopOI,prepost, sep="-"))) %>%
           colSums(na.rm = TRUE) %>%
           as.data.frame() %>% `colnames<-`(c("totalRank")) %>%
           mutate(avgRank = totalRank / nrankOI))
}

ui <- dashboardPage(
  # Dashboard Header ----------------------------------------------------
  dashboardHeader(
    title = "ITN Dashboard"
  ),
  # Dashboard Sidebar ----------------------------------------------------
  dashboardSidebar(
    sidebarMenu(
      menuItem("Courses", tabName = "tab_courses", icon = icon("chalkboard"),
               menuSubItem("Plots", tabName = "tab_courses_plots"),
               menuSubItem("Tables", tabName = "tab_courses_tables")),
      menuItem("Workshops", tabName = "tab_workshops", icon = icon("people-group")),
      menuItem("Software Usage", tabName = "tab_software_usage", icon = icon("robot")),
      menuItem("Collaborations", tabName = "tab_collabs", icon = icon("people-arrows")),
      menuItem("About", tabName = "tab_about", icon = icon("info"))
    )
  ),
  # Dashboard Body ----------------------------------------------------
  dashboardBody(
    # Custom style
    includeCSS("www/style.css"),
    # Favicon
    tags$head(tags$link(rel="shortcut icon", href="i/img/favicon.ico")),

    tabItems(
      # Courses Tab ----------------------------------------------------
      tabItem(tabName = "tab_courses_plots",
              # First row
              fluidRow(
                box(title = "Visitors across Educational Resources",
                    plotOutput("plot_visitor_website")),

                box(title = "Engagement by Modality",
                    selectInput("modality", "Modality:",
                                c("Website" = "website",
                                  "Leanpub" = "leanpub",
                                  "Coursera" = "coursera")),
                    plotOutput("plot_engagement_modality"))
              ),
              # Second row
              fluidRow(
                box(title = "Website Engagement",
                    selectInput("metric", "Metric:",
                                c("Screen Page Views per User" = "screen_page_views_per_user",
                                  "Average Session Duration" = "average_session_duration",
                                  "Event Count per User" = "event_count_per_user",
                                  "Engagement Rate" = "engagement_rate")),
                    plotOutput("plot_engagement_website")),

                box(title = "Course Engagement by Target Audience",
                    plotOutput("plot_engagement_target"))
              ),
              # Third row
              fluidRow(
                box(title = "Total Number of Learners for each Course",
                    plotOutput("plot_learner_course")),

                box(title = "Number of Coursera Enrollments by Course",
                    plotOutput("plot_coursera"),
                    footer = "*Data was manually updated, and may be outdated.")
              ),
              # Fourth row
              fluidRow(
                box(title = "Number of Leanpub Enrollments by Course",
                    plotOutput("plot_leanpub"),
                    footer = "*Data was manually updated, and may be outdated."),

                box(title = "Course Popularity over Time",
                    plotOutput("plot_learner_launch_date"))
              )
      ),
      # Tables Tab ----------------------------------------------------
      tabItem(tabName = "tab_courses_tables",
              tabBox(side = "left",
                     height = "400px",
                     tabPanel("User Totals",
                              DTOutput("table_user_total")),
                     tabPanel("User Engagement",
                              DTOutput("table_user_engagement")),
                     width = 12
              )

      ),

      # Workshops Tab ----------------------------------------------------
      tabItem(tabName = "tab_workshops",
              # First Row
              fluidRow(
                box(title = "Workshop Recommendation Likelihood",
                    plotOutput("plot_workshop_recommendation"),
                    footer = textOutput("percent_rec")),

                box(title = "Workshop Relevance",
                    plotOutput("plot_workshop_relevance"))
              ),
              # Second Row
              fluidRow(
                box(title = "Change in confidence on workshop topic(s)",
                    plotOutput("plot_workshop_confidence_pooled")),
                box(title = "Change in confidence for specific workshops",
                    plotOutput("plot_workshop_confidence"))
              ),
              # Third Row
              fluidRow(
                box(title = "Workshop Registrant Career Stage",
                    width = 12,
                    plotOutput("plot_workshop_career_stage"))
              ),
              # Second Row
              fluidRow(
                box(title = "Workshop Review: What Did You Like Most?",
                    plotOutput("plot_workshop_review")),
                box(title = "Workshop Recommendations for Improvement",
                    plotOutput("plot_workshop_improvement"))
              )
      ),
      # Software Usage Tab ----------------------------------------------------
      tabItem(tabName = "tab_software_usage",
              # First Row
              fluidRow(
                infoBox("Loqui",
                        "A Shiny app for Creating Automated Videos",
                        href = "https://loqui.fredhutch.org/",
                        icon = icon("video")),
                infoBox("Number of Loqui Users", 20, icon = icon("user-group")),
                infoBox("Number of Videos made with Loqui", 490, icon = icon("play"))
              ),
              # Second Row
              fluidRow(
                box(title = "Monthly CRAN Downloads",
                    width = 12,
                    plotOutput("plot_monthly_cran_download"),
                    footer = "*Dashed vertical lines denote when software was published on CRAN.")
              ),
              # Third Row
              fluidRow(
                tabBox(side = "left",
                     height = "300px",
                     tabPanel("Package Download Totals",
                              DTOutput("table_cran_downloads")),
                     tabPanel("Package Download Totals - Last 6 months",
                              DTOutput("table_cran_sixmonths")),
                     width = 8
                )
              )
      ),
      # Collaborations Tab ----------------------------------------------------
      tabItem(tabName = "tab_collabs",
              # First Row
              fluidRow(
                box(title = "All Collaborations",
                    plotOutput("plot_collaboration_all")),
                box(title = "ITCR-Only Collaborations",
                    plotOutput("plot_collaboration_itcr"))
              ),
              #Second Row
              fluidRow(
                box(title = "All Collaborations: Specifying Affiliation",
                    plotOutput("plot_collaboration_all_color"))
              )
      ),
      # About Tab ----------------------------------------------------
      tabItem(tabName = "tab_about",
              "UNDER CONSTRUCTION"
      )

    )
  )
)

# Server ----------------------------------------------------
server <- function(input, output) {
  # Data: ITCR Course ----------------------------------------------------
  itcr_course_data_raw <- reactiveFileReader(time_interval,
                                             NULL,
                                             "https://raw.githubusercontent.com/FredHutch/itn-dashboard/main/data/itcr_course_metrics.csv",
                                             readr::read_csv)

  itcr_course_data <- reactive({
    itcr_course_data <- itcr_course_data_raw() %>%
      mutate(target_audience = replace_na(target_audience, "Everyone"))

    itcr_course_data$webAndEnrollmentTotals <- itcr_course_data %>%
      select(website_count, coursera_count, leanpub_count) %>%
      rowSums(na.rm = TRUE)

    itcr_course_data
  })

  itcr_course_data_long <- reactive({
    itcr_course_data() %>%
      select(c("website",
               "totalUsers",
               "coursera_count",
               "leanpub_count",
               "target_audience")) %>%
      tidyr::pivot_longer(!c(website, target_audience),
                          names_to = "modality",
                          values_to = "learner_count") %>%
      filter(!(website %in% c("ITN Website", "OTTR website", "metricminer.org"))) %>%
      mutate(modality = case_when(
        modality == "leanpub_count" ~ "Total Leanpub Enrollments",
        modality == "coursera_count" ~ "Total Coursera Enrollments",
        modality == "totalUsers" ~ "Website Learners",
        TRUE ~ modality
      ))
  })

  # Data: ITCR Google Analytics ----------------------------------------------------
  ga_metrics <-  reactiveFileReader(time_interval,
                                    NULL,
                                    "https://raw.githubusercontent.com/FredHutch/itn-dashboard/main/data/itcr_ga_metric_data.csv",
                                    readr::read_csv)

  user_totals <- reactive({
    ga_metrics() %>%
      clean_names() %>%
      select(website, active_users, average_session_duration) %>%
      mutate(average_session_duration = round(average_session_duration, digits = 0))
  })

  user_engagement <- reactive({
    ga_metrics() %>%
      clean_names() %>%
      select(website, screen_page_views_per_user,
             sessions, screen_page_views, engagement_rate) %>%
      mutate(screen_page_views_per_user = round(screen_page_views_per_user, 0),
             engagement_rate = round(engagement_rate, 2))
  })


  # Data: Course Engagement by Modality ----------------------------------------------------
  course_raw <- reactiveFileReader(time_interval,
                                   NULL,
                                   "https://docs.google.com/spreadsheets/d/1-8vox2LzkVKzhmSFXCWjwt3jFtK-wHibRAq2fqbxEyo/edit?usp=sharing",
                                   googlesheets4::read_sheet,
                                   sheet = "Course_data")


  course_processed <- reactive({
    course_raw() %>%
      pivot_longer(cols = contains("count"), names_to = "modality", values_to = "number_of_learners") %>%
      mutate(course_name = factor(website)) %>%
      separate(modality, sep = "_", into = c("modality", "meh")) %>%
      mutate(modality = factor(modality, levels = c("website", "leanpub", "coursera"),
                               labels = c("website", "leanpub", "coursera"))) %>%
      mutate(course_order = case_when(target_audience == "Leadership" ~ 1,
                                      target_audience == "New to data" ~ 2,
                                      target_audience == "Software developers" ~ 3)) %>%
      rename("Target Audience" =  target_audience) %>%
      filter(modality == input$modality)

  })

  # Data: ITCR Slido ----------------------------------------------------
  itcr_slido_data <- reactiveFileReader(time_interval,
                                        NULL,
                                        "https://raw.githubusercontent.com/FredHutch/itn-dashboard/main/data/itcr_slido_data.csv",
                                        readr::read_csv)
  
  itcr_slido_data_processed <- reactive({  
    itcr_slido_data() %>% clean_names() %>% 
    mutate(participant = coalesce(user_id, participant_id)) %>%
    filter(!is.na(participant))
  })
  
  itcr_slido_data_rec <- reactive({
    itcr_slido_data_processed() %>% 
      filter(!str_detect(event_name, "Pre")) %>% #remove pre workshop survey that doesn't ask the relevant question
      select(matches("recommend_")) %>% 
      mutate(merged_likely_rec = as.integer(coacross(everything())))
  })
  
  
  output$percent_rec <- renderText({ 
    paste0(round(sum(itcr_slido_data_rec()$merged_likely_rec >= 8, na.rm = TRUE) / sum(!is.na(itcr_slido_data_rec()$merged_likely_rec)) * 100, digits=1),  
           " % of responses rated their recommendation likelihood as an 8 or higher.")
  })
  

  # Data: Workshop Registrant Career Stage ----------------------------------------------------
  career_stage_counts_raw <- reactiveFileReader(time_interval,
                                                NULL,
                                                "https://docs.google.com/spreadsheets/d/1-8vox2LzkVKzhmSFXCWjwt3jFtK-wHibRAq2fqbxEyo/edit?usp=sharing",
                                                googlesheets4::read_sheet,
                                                sheet = "Workshop attendee type totals")
  
  # This data is manually curated and does NOT automatically update
  #This googlesheet has the workshop name in the first column and has counts for a specific career stage in each column.
  #The final row of the google sheet is the total (column sum)

  career_stage_counts_summed <- reactive({
    #tmp <- career_stage_counts_raw() %>%
      #select(-1) %>% #drop the first column
      #slice(1:(n() - 1)) #drop the last row

    #colSums(tmp) #colSum to get the values that were in the last row....
    
    if(tolower(career_stage_counts_raw() %>% slice(n()) %>% select(1)) == "total"){
      t(career_stage_counts_raw() %>% slice(n()) %>% select(-1))
    }
    
  })

  career_stage_processed <- reactive({
    career_stage_processed <- data.frame(
      Stage = rownames(career_stage_counts_summed()),
      count = as.numeric(career_stage_counts_summed()[,1]),
      stringsAsFactors = FALSE
    )

    career_stage_processed$Trainee <- ifelse(career_stage_processed$Stage %in% c("Phd student",
                                                                                 "postdoc",
                                                                                 "Master's student",
                                                                                 "Research tech",
                                                                                 "undergrad"),
                                             "yes",
                                             "no")
    career_stage_processed
  })

  # Data: Poll Results ---------------------------------------------------
  poll_data <- reactive({
    itcr_slido_data() %>%
      clean_names() %>%
      mutate(merged_likely_rec = if_else(is.na(how_likely_would_you_be_to_recommend_this_workshop), how_likely_would_you_be_to_recommend_this_workshop_2,
                                         how_likely_would_you_be_to_recommend_this_workshop))
  })


  poll_results <- reactive({
    udpipe::udpipe_annotate(ud_model, x = poll_data()$what_did_you_like_most_about_the_workshop) %>%
      as.data.frame() %>%
      dplyr::filter(upos %in% c("NOUN", "ADJ", "ADV")) %>%
      mutate(lemma= tolower(lemma)) %>%
      count(lemma)
  })

  poll_rec_results <- reactive({
    udpipe::udpipe_annotate(ud_model, x = poll_data()$please_share_any_recommendations_you_have_for_improvements) %>%
      as.data.frame() %>%
      filter(upos %in% c("NOUN", "ADJ", "ADV")) %>%
      mutate(lemma= tolower(lemma)) %>%
      count(lemma)
  })

  # Data: CRAN Downloads ----------------------------------------------------
  cran_download <-  reactiveFileReader(time_interval,
                                       NULL,
                                       "https://raw.githubusercontent.com/FredHutch/itn-dashboard/main/data/cran_download_stats.csv",
                                       readr::read_csv)

  # Data: Collaborations ----------------------------------------------------
  collabs_raw <-  reactiveFileReader(time_interval,
                                     NULL,
                                     "https://docs.google.com/spreadsheets/d/1-8vox2LzkVKzhmSFXCWjwt3jFtK-wHibRAq2fqbxEyo/edit?usp=sharing",
                                     googlesheets4::read_sheet)

  collabs_processed <- reactive({
    collabs_raw() %>%
      separate_rows("Category", sep = ", ", ) %>%
      mutate(Category = trimws(Category)) %>%
      filter(Category != "?")
  })

  # Table: User Totals ----------------------------------------------------
  output$table_user_total <- renderDT({
    DT::datatable(
      user_totals(),
      colnames = c("Website", "Active Users", "Avg Session Duration"),
      options = list(lengthChange = FALSE, # remove "Show X entries"
                     searching = FALSE,
                     scrollY = "450px"), # remove Search box
      # For the table to grow/shrink
      fillContainer = TRUE,
      escape = FALSE
    )
  })

  # Table: User Engagement ----------------------------------------------------
  output$table_user_engagement <- renderDT({
    DT::datatable(
      user_engagement(),
      colnames = c("Website", "Screen Page Views per User", "Sessions",
                   "Screen Page Views", "Engagement Rate"),
      options = list(lengthChange = FALSE, # remove "Show X entries"
                     searching = FALSE,
                     scrollY = "450px"), # remove Search box
      # For the table to grow/shrink
      fillContainer = TRUE,
      escape = FALSE
    )
  })




  # Plot: Visitors across Educational Resources ----------------------------------------------------
  output$plot_visitor_website <- renderPlot({
    itcr_course_data() %>%
      # Filter out ITN Website since it is not an "Educational Resource"
      filter(!(website %in% c("ITN Website", "widget", "DaSL Collection",
                              "proof", "metricminer.org", "OTTR website",
                              "Developing_WDL_Workflows", "Using Leanpub Course",
                              "ocdo.fredhutch.org"))) %>%
      ggplot(aes(x = reorder(website, -totalUsers), y = totalUsers, fill = target_audience)) +
      geom_bar(stat = "identity") +
      geom_text(aes(label = totalUsers), hjust = 1.05,
                colour = "lightgray", fontface = "bold") +
      theme_classic() +
      coord_flip() +
      theme(text = element_text(size = 17, family = "Arial"),
            legend.position = "inside",
            legend.position.inside = c(0.75,0.75)) +
      labs(x = NULL,
           y = "Number of Visitors",
           fill = "Target Audience") +
      scale_fill_manual(values=cbLEPalette) +
      guides(fill=guide_legend(nrow=3))
  })

  # Plot: Engagement by Modality ----------------------------------------------------
  output$plot_engagement_modality <- renderPlot({
    course_processed()  %>%
      filter(modality == input$modality) %>%
      # Some courses have 0 learners
      filter(number_of_learners > 0) %>%
      ggplot(aes(x = fct_reorder(course_name, number_of_learners),
                 y = number_of_learners, fill = `Target Audience`)) +
      geom_col() +
      geom_text(aes(label = number_of_learners), hjust = 1.05,
                colour = "lightgray", fontface = "bold") +
      coord_flip() +
      scale_fill_manual(values=cbLEPalette) +
      theme_classic() +
      theme(legend.position = "inside",
            legend.position.inside = c(0.75, 0.25),
            text = element_text(size = 17, family = "Arial")) +
      labs(x = NULL,
           y = "Number of Learners") +
      guides(fill=guide_legend(nrow=3))
  })

  # Plot: Website Engagement ----------------------------------------------------
  output$plot_engagement_website <- renderPlot({
    itcr_course_data() %>%
      clean_names() %>%
      select(website, screen_page_views_per_user, average_session_duration,
             event_count_per_user, engagement_rate, target_audience) %>%
      pivot_longer(!c(website, target_audience),
                   names_to = "metric_name",
                   values_to = "value") %>%
      filter(!(website %in% c("ITN Website", "OTTR website", "metricminer.org")),
             metric_name == input$metric) %>%
      ggplot(aes(x = website, y = value, fill = target_audience)) +
      geom_bar(position = "dodge", stat = "identity") +
      geom_text(aes(label = round(value, 1), hjust = 1.05),
                colour = "lightgray", fontface = "bold") +
      coord_flip() +
      theme_classic() +
      labs(x = NULL,
           y = NULL,
           fill = "Target Audience") +
      scale_fill_manual(values=cbLEPalette, limits = c("Leadership", "New to data", "Software developers")) +
      scale_x_discrete(limits = c("Leadership in Cancer Informatics", "NIH Data Sharing", "Ethical Data Handling", "Overleaf and Latex for Scientific Articles", "AI for Decision Makers",
                                  "Reproducibility in Cancer Informatics", "Choosing Genomics Tools", "Computing for Cancer Informatics",
                                  "Documentation and Usability", "Advanced Reproducibility", "AI for Efficient Programming", "GitHub Automation for Scientists")) +
      theme(text = element_text(size = 17, family = "Arial"),
            legend.position = "bottom") +
      guides(fill=guide_legend(nrow=3))
  })

  # Plot: Course Engagement by Target Audience ----------------------------------------------------
  output$plot_engagement_target <- renderPlot({
    itcr_course_data_long() %>%
      filter(!(website %in% c("widget", "DaSL Collection", "Developing_WDL_Workflows", "proof", "Using Leanpub Course", "ocdo.fredhutch.org"))) %>%
      group_by(modality, target_audience) %>%
      summarize(total_learners = sum(learner_count, na.rm = TRUE)) %>%
      ggplot(aes(x = reorder(modality, -total_learners), y = total_learners, fill = target_audience)) +
      geom_bar(stat = "identity", na.rm = TRUE) +
      geom_text(aes(label = total_learners), hjust = 1.055, na.rm = TRUE,
                colour = "lightgray", fontface = "bold") +
      coord_flip() +
      theme_classic() +
      theme(legend.position = "none",
            text = element_text(size = 17, family = "Arial")) +
      labs(x = NULL,
           y = "Visitors/Enrollees",
           fill = "Target Audience") +
      #ylim(c(0, 4200)) +
      facet_wrap(~target_audience, nrow=3) +
      scale_fill_manual(values=cbLEPalette, limits = c("Leadership", "New to data", "Software developers"))
  })    

  # Plot: Learner by Course ----------------------------------------------------
  output$plot_learner_course <- renderPlot({
    itcr_course_data_long() %>%
      group_by(website, target_audience) %>%
      summarize(total_learners = sum(learner_count, na.rm = TRUE)) %>%
      filter(!(website %in% c("widget", "DaSL Collection", "Developing_WDL_Workflows", "proof", "Using Leanpub Course", "ocdo.fredhutch.org"))) %>%
      ggplot(aes(x = reorder(website, -total_learners), y = total_learners, fill = target_audience)) +
      geom_bar(stat = "identity") +
      geom_text(aes(label = total_learners), hjust = 1.05, na.rm = TRUE,
                colour="lightgray", fontface = "bold") +
      labs(x = NULL,
           y = "Total Learners by Course",
           fill = "Target Audience",
           title = NULL) +
      coord_flip() +
      theme_classic() +
      theme(text = element_text(size = 17, family = "Arial"),
            legend.position = "inside",
            legend.position.inside = c(0.75, 0.75)) +
      scale_fill_manual(values=cbLEPalette) +
      guides(fill=guide_legend(nrow=3))
  })

  # Plot: Coursera Learners ----------------------------------------------------
  output$plot_coursera <- renderPlot({
    itcr_course_data() %>%
      filter(coursera_count > 0) %>%
      ggplot(aes(x = reorder(website, -coursera_count), y = coursera_count, fill = target_audience)) +
      geom_bar(stat = "identity", na.rm = TRUE) +
      geom_text(aes(label = coursera_count), hjust = 1.05, na.rm = TRUE,
                colour = "lightgray", fontface = "bold") +
      coord_flip() +
      theme_classic() +
      labs(x = NULL,
           y = "Coursera enrollments",
           fill = "Target Audience",
           title = NULL) +
      ylim(c(0, 1200)) +
      scale_fill_manual(values = cbLEPalette) +
      theme(text = element_text(size = 17, family = "Arial"),
            legend.position = "inside",
            legend.position.inside = c(0.75, 0.75)) +
      guides(fill=guide_legend(nrow=3))
  })

  # Plot: Leanpub Learners ----------------------------------------------------
  output$plot_leanpub <- renderPlot({
    itcr_course_data() %>%
      filter(leanpub_count > 0) %>%
      ggplot(aes(x = reorder(website, -leanpub_count), y = leanpub_count, fill = target_audience)) +
      geom_bar(stat = "identity", na.rm = TRUE) +
      geom_text(aes(label = leanpub_count), hjust = 1.05, na.rm = TRUE,
                colour = "lightgray", fontface = "bold") +
      coord_flip() +
      theme_classic() +
      labs(x = NULL,
           y = "Leanpub enrollments",
           fill = "Target Audience",
           title = NULL) +
      ylim(c(0, 40)) +
      scale_fill_manual(values = cbLEPalette) +
      theme(text = element_text(size = 17, family = "Arial"),
            legend.position = "inside",
            legend.position.inside = c(0.75, 0.75)) +
      guides(fill=guide_legend(nrow=3))
  })

  # Plot: Learners by Launch Date ----------------------------------------------------
  output$plot_learner_launch_date <- renderPlot({
    itcr_course_data() %>%
      filter(!(website %in% c("ITN Website", "OTTR website", "metricminer.org"))) %>%
      mutate(duration = today() - website_launch) %>%
      ggplot(aes(x = duration, y = website_count, color = target_audience)) +
      geom_point() +
      theme(panel.grid = element_line("black", linewidth = 0.25),
            panel.background = element_blank(),
            panel.border = element_rect("black", fill=NA, linewidth=0.5),
            legend.position = "bottom",
            text = element_text(size = 17, family = "Arial")) +
      labs(x = "How long the course has been out (days)",
           y = "Bookdown Views",
           color = "Target Audience",
           title = NULL) +
      scale_color_manual(values=cbLEPalette, limits = c("Leadership", "New to data", "Software developers")) +
      ggrepel::geom_text_repel(
        aes(label = website),
        family = "Poppins",
        size = 3.5,
        min.segment.length = 0, 
        seed = 42, 
        box.padding = 0.5,
        max.overlaps = Inf,
        arrow = arrow(length = unit(0.010, "npc")),
        nudge_x = .15,
        nudge_y = .5,
        color = "black"
      )
  })

  
  
  # Plot: Recommendation likelihood ------------------------------------------------------
  
  output$plot_workshop_recommendation <- renderPlot({
    itcr_slido_data_processed() %>%
      filter(!str_detect(event_name, "Pre")) %>% #remove pre workshop survey that doesn't ask the relevant question
      select(matches("recommend_")) %>% 
      mutate(merged_likely_rec = as.integer(coacross(everything()))) %>%
      ggplot(aes(merged_likely_rec)) +
      geom_bar(fill = "#28ae80") +
      theme_classic() +
      scale_x_continuous(breaks= c(1:10), labels=c(1:10), limits=c(0.5,10.5)) +
      coord_cartesian(clip="off") +
      geom_text(stat = "count", aes(label = after_stat(count)), vjust= 1.4,
                colour = "lightgray", fontface = "bold") +
      theme(text = element_text(size = 17, family = "Arial")) +
      labs(y = "Count",
           x = "Rating") +
      ggtitle("How likely are you to recommend this workshop?") +
      annotation_custom(textGrob("Most\nLikely", gp=gpar(fontsize=8, fontface = "bold")),xmin=10,xmax=10,ymin=-13,ymax=-13) +
      annotation_custom(textGrob("Least\nLikely", gp=gpar(fontsize=8, fontface= "bold")),xmin=1,xmax=1,ymin=-13,ymax=-13)
    
  })
  
  
      
  # Plot: Workshop Relevance (e.g., Positive impact likelihood) ---------------------------------------------------
  
  output$plot_workshop_relevance <- renderPlot({
    itcr_slido_data_processed() %>%
      filter(!str_detect(event_name, "Pre")) %>% #remove pre workshop survey that doesn't ask the relevant question
      select(matches("relevant|positive_impact")) %>%
      select(!matches("current_research_work")) %>%
      mutate(merged_relevant_likely = as.integer(coacross(everything()))) %>%
      ggplot(aes(merged_relevant_likely)) +
      geom_bar(fill = "#28ae80") +
      theme_classic() +
      scale_x_continuous(breaks= c(1:10), labels=c(1:10), limits=c(0.5,10.5)) +
      coord_cartesian(clip="off") +
      geom_text(stat = "count", aes(label = after_stat(count)), vjust= 1.4,
                colour = "lightgray", fontface = "bold") +
      theme(text = element_text(size = 17, family = "Arial")) +
      labs(y = "Count",
           x = "Rating") +
      ggtitle("How likely is this workshop to have\na positive impact on your work?") +
      annotation_custom(textGrob("Most\nLikely", gp=gpar(fontsize=8, fontface = "bold")),xmin=10,xmax=10,ymin=-5.5,ymax=-5.5) +
      annotation_custom(textGrob("Least\nLikely", gp=gpar(fontsize=8, fontface= "bold")),xmin=1,xmax=1,ymin=-5.5,ymax=-5.5)
      
    
  })
  
  # Plot: Pre- and post- workshop confidence (pooled) ----------------------------------------
  output$plot_workshop_confidence_pooled <- renderPlot({
    rbind(
      itcr_slido_data_processed() %>%  
        filter(!str_detect(event_name, "Q2-NIH_")) %>%
        select(event_name, contains("confident")) %>%
        mutate(across(contains("confident"), as.character)) %>% #getting an error about combining characters and doubles, so making everything a character before changing to an integer later after dropping irrelevant characters
        pivot_longer(contains("confident"), values_to = "value", names_to = "question") %>%
        mutate(pre_post = if_else(grepl("now", question), "post", "pre")) %>% 
        filter(str_length(value) <= 2) %>% #filter out the ones that are phrases and not numbers
        filter(!str_detect(event_name, "GLBIO")) %>% #max ratings of 5 so filter out
        select(value, pre_post),
      itcr_slido_data_processed() %>%
        filter(event_name == "Q2-NIH_PreSurvey") %>%
        select(contains("confident")) %>% 
        mutate(across(contains("confident"), as.character)) %>% #getting an error about combining characters and doubles, so making everything a character before changing to an integer later after dropping irrelevant characters
        pivot_longer(everything(), values_to = "value", names_to = "question") %>%
        mutate(pre_post = "pre") %>%
        select(value, pre_post),
      itcr_slido_data_processed() %>%
        filter(str_detect(event_name, "Q2-NIH_") & event_name != "Q2-NIH_PreSurvey") %>%
        select(contains("confident")) %>%
        mutate(across(contains("confident"), as.character)) %>% #getting an error about combining characters and doubles, so making everything a character before changing to an integer later after dropping irrelevant characters
        pivot_longer(everything(), values_to = "value", names_to = "question") %>%
        mutate(pre_post = "post") %>%
        select(value, pre_post)
    ) %>% 
      drop_na() %>%
      mutate(value = as.integer(value),
             pre_post = factor(pre_post, levels = c("pre", "post"))
      ) %>%
      ggplot(aes(x = value, y=pre_post, fill=pre_post)) +
      geom_boxplot(outliers = FALSE) +
      geom_jitter(aes(fill=pre_post), height=0.1, width=0.35, alpha=0.4, size=1.5, shape=21, color="black", stroke=1.5) +
      theme_bw() + theme(panel.background = element_blank()) +
      theme(legend.position = "bottom") +
      xlab("Confidence Rank") + scale_x_continuous(breaks = 1:10, labels = 1:10) +
      ylab("") +
      ggtitle("How confident do you feel about ...") +
      scale_fill_discrete(name = "Pre or post workshop?") +
      coord_cartesian(clip = 'off') +
      annotation_custom(textGrob("Most\nConfident", gp=gpar(fontsize=8, fontface = "bold")),xmin=10,xmax=10,ymin=0.1,ymax=0.1) +
      annotation_custom(textGrob("Least\nConfident", gp=gpar(fontsize=8, fontface= "bold")),xmin=1,xmax=1,ymin=0.1,ymax=0.1) +
      theme(axis.text.y=element_blank(),
            axis.ticks.y=element_blank(),
            text = element_text(size = 17, family = "Arial"))
  })
  
  # Plot: Pre- and post- workshop confidence (workshop specific) ----------------------------
  output$plot_workshop_confidence <- renderPlot({
    
  })
  
  # Plot: Workshop Career Stage ----------------------------------------------------
  output$plot_workshop_career_stage <- renderPlot({
    career_stage_processed() %>%
      ggplot(aes(x = reorder(Stage, -count), y = count, fill = Trainee)) +
      geom_bar(stat = "identity") +
      geom_text(aes(label = count), hjust = 1.2,
                colour ="lightgray", fontface = "bold") +
      coord_flip() +
      xlab(NULL) +
      ylab("Number of Registrants") +
      theme_bw() +
      theme(panel.background = element_blank(),
            panel.grid = element_blank(),
            text = element_text(size = 17, family = "Arial")) +
      scale_fill_manual(values = c("#440154", "#28ae80"))
  })

  # Plot: Workshop Review ----------------------------------------------------
  output$plot_workshop_review <- renderPlot({
    wordcloud::wordcloud(words = poll_results()$lemma,
                         freq = poll_results()$n,
                         colors = c("#98fb98", "#83D475", "#355E3B"),
                         min.freq = 3, scale = c(3, .4))
  })

  # Plot: Workshop Feedback ----------------------------------------------------
  output$plot_workshop_improvement <- renderPlot({
    wordcloud::wordcloud(words = poll_rec_results()$lemma,
                         freq= poll_rec_results()$n,
                         colors = c("#98fb98", "#83D475", "#355E3B"),
                         min.freq = 3, scale = c(4, .4))
  })

  # Plot: Monthly CRAN Download ----------------------------------------------------
  output$plot_monthly_cran_download <- renderPlot({
    cran_download() %>%
      ggplot(aes(Month, monthly_downloads, group=package, color = package)) +
      geom_line() +
      geom_point() +
      scale_colour_manual(values=viridis_cc) +
      theme(panel.background = element_blank(),
            panel.grid = element_blank(),
            text = element_text(size = 17, family = "Arial")) +
      geom_vline(aes(xintercept = "2019-05"), linetype='dashed', color = '#addc30') + #text2speech published date
      geom_vline(aes(xintercept="2022-02"), linetype='dashed', color = '#28ae80') + #ottrpal published date
      geom_vline(aes(xintercept="2023-07"), linetype='dashed', color = '#2c728e') + #conrad published date
      geom_vline(aes(xintercept="2024-02"), linetype="dashed", color = '#fde725') + #metricminer published date
      theme(axis.text.x = element_text(angle = 90),
            legend.position = "bottom") + #clean up x-axis labels
      labs(x = NULL,
           y = "Monthly Downloads",
           color = "R Packages")
  })

  # Table: Total CRAN downloads by package
  output$table_cran_downloads <- renderDT({
    DT::datatable(
      cran_download() %>%
      group_by(package) %>%
      summarize(total_downloads = sum(monthly_downloads)),
      colnames = c("Package", "Total Downloads"),
      options = list(lengthChange = FALSE, # remove "Show X entries"
                     searching = FALSE,
                     scrollY = "150px"), # remove Search box
      # For the table to grow/shrink
      fillContainer = TRUE,
      escape = FALSE
    )
  })
  
  # Table: Total CRAN downloads by package last 6 months
  output$table_cran_sixmonths <- renderDT({
    DT::datatable(
      cran_download() %>%
        filter(Month > format(as.Date(today() - months(6)), "%Y-%m")) %>% #within 6 months
        group_by(package) %>%
        summarize(total_downloads = sum(monthly_downloads)),
      colnames = c("Package", "Total Downloads"),
      options = list(lengthChange = FALSE, # remove "Show X entries"
                     searching = FALSE,
                     scrollY = "150px"), # remove Search box
      # For the table to grow/shrink
      fillContainer = TRUE,
      escape = FALSE
    )
  })
  
  # Plot: All Collaborations ----------------------------------------------------
  output$plot_collaboration_all <- renderPlot({
    collabs_processed() %>%
      count(Category) %>%
      ggplot(aes(y = n, x = reorder(Category,-n), fill = Category)) +
      geom_bar(position = "dodge", stat = "identity") +
      geom_text(aes(label = n), hjust = 1.05,
                colour = "white", fontface = "bold") +
      coord_flip() +
      theme_classic() +
      theme(strip.text.x = element_text(size = 6),
            legend.position="none",
            text = element_text(size = 17, family = "Arial"),
            plot.margin = unit(c(.75,.5,.5,.5), "cm")) +
      xlab(NULL) +
      ylab(NULL)
  })

  # Plot: ITCR Collaborations ----------------------------------------------------
  output$plot_collaboration_itcr <- renderPlot({
    collabs_processed() %>%
      filter(ITN_ITCR_or_external == "ITCR") %>%
      count(Category) %>%
      ggplot(aes(y =n, x=reorder(Category, -n), fill = Category )) +
      geom_bar(position="dodge", stat = "identity") +
      geom_text(aes(label = n), hjust = 1.05,
                colour = "white", fontface = "bold") +
      coord_flip() +
      theme_classic() +
      theme(text = element_text(size = 17, family = "Arial"),
            legend.position = "none",
            plot.margin = unit(c(.75,.5,.5,.5), "cm")) +
      xlab(NULL) +
      ylab(NULL)
  })
  
  # Plot: All Collaborations colored by ITCR or not ------------------------------
  output$plot_collaboration_all_color <- renderPlot({
    collabs_processed() %>%
      group_by(Category, ITN_ITCR_or_external) %>%
      mutate(ITN_ITCR_or_external = 
               factor(case_when(ITN_ITCR_or_external == "external" ~ "Not ITCR",
                                ITN_ITCR_or_external == "external (was after leaving)" ~ "Not ITCR",
                                ITN_ITCR_or_external == "external NIH/NCI intermural" ~ "Not ITCR",
                                ITN_ITCR_or_external == "ITCR" ~ "ITCR",
                                ITN_ITCR_or_external == "ITN" ~ "ITN",
                                ITN_ITCR_or_external == "NCI" ~ "Not ITCR",
                                ITN_ITCR_or_external == "neither" ~ "Not ITCR",
                                ITN_ITCR_or_external == "NIH" ~ "Not ITCR"),
                      levels = rev(c("ITCR", "Not ITCR", "ITN")))) %>%
      summarize(n= n()) %>%
      ggplot(aes(y = n, x = reorder(Category,-n, sum), fill = ITN_ITCR_or_external)) +
      geom_bar(position = "stack", stat = "identity") +
      geom_text(aes(label = after_stat(y), group = Category), hjust = 1.05,
                stat = 'summary', fun = sum,
                colour = "white", fontface = "bold") +
      coord_flip() +
      theme_classic() +
      labs(fill = "ITCR Collab?") +
      theme(strip.text.x = element_text(size = 6),
            legend.position="inside",
            legend.position.inside = c(0.75, 0.75),
            text = element_text(size = 17, family = "Arial"),
            plot.margin = unit(c(.75,.5,.5,.5), "cm")) + 
      xlab(NULL) +
      ylab(NULL) +
      scale_fill_discrete(na.translate = F)
  })


}

# Code for Deployment to Hutch servers
addResourcePath("/i", file.path(getwd(), "www"))
options <- list()
if (!interactive()) {
  options$port = 3838
  options$launch.browser = FALSE
  options$host = "0.0.0.0"

}

shinyApp(ui, server, options=options)