Week 8: R Shiny - Interactive Dashboards

The USP skill that gets EDUSHARK graduates hired

Duration: 7 Days | Level: Core Module | Prereq: Weeks 6-7

Power BI shows data. R Shiny builds decision tools. This week is the centrepiece of the EDUSHARK programme. You will build a complete interactive dashboard from a single R file, learn how reactivity works under the hood, and deploy your app to a public URL on shinyapps.io. By Friday, you have something you can pin to your portfolio.

What You'll Learn This Week

Day 1: Hello, Shiny
Day 2: UI Components and Layout
Day 3: Reactivity Deep Dive
Day 4: Interactive Visualisations (plotly, DT)
Day 5: shinydashboard and bslib
Day 6: Deployment
Day 7: Project - Interactive Business Dashboard

Day 1

Hello, Shiny

A complete interactive app in under 30 lines

A Shiny app has two pieces: a ui that describes the page, and a server function that reads inputs and writes outputs. Shiny re-runs the server logic automatically whenever an input changes - this is reactivity, and it is the whole magic.

library(shiny); library(ggplot2)

ui <- fluidPage(
  titlePanel("Hello Shiny"),
  sidebarLayout(
    sidebarPanel(sliderInput("bins", "Number of bins:", 5, 50, 30)),
    mainPanel(plotOutput("hist"))
  )
)

server <- function(input, output, session) {
  output$hist <- renderPlot({
    ggplot(faithful, aes(eruptions)) +
      geom_histogram(bins = input$bins, fill = "#0077B6", colour = "white") +
      theme_minimal()
  })
}

shinyApp(ui, server)
                    

Save as app.R and click Run App in RStudio.

Day 2

UI Components and Layout

Inputs you'll use weekly

textInput, numericInput, sliderInput, selectInput, checkboxInput, checkboxGroupInput, radioButtons, dateInput, dateRangeInput, fileInput, actionButton.

Outputs and their renderers

plotOutput ↔ renderPlot, tableOutput ↔ renderTable, DT::DTOutput ↔ DT::renderDT, textOutput ↔ renderText, uiOutput ↔ renderUI.

Layouts

fluidPage, sidebarLayout, fluidRow + column, tabsetPanel, navbarPage.

Day 3

Reactivity Deep Dive

The mental model you must own

Sources

Inputs: input$x. The user clicks; this is where re-evaluation starts.

Conductors

reactive({}) expressions. Compute values used by multiple outputs.

Endpoints

Outputs (output$plot) and observers. Where reactivity ends.

server <- function(input, output, session) {

  filtered <- reactive({
    orders %>%
      filter(region == input$region,
             order_date %within% input$dr[1] %--% input$dr[2])
  })

  output$kpi   <- renderText(sum(filtered()$amount))
  output$chart <- renderPlot(ggplot(filtered(), aes(...)) + ...)
  output$table <- DT::renderDT(filtered())
}
                    

Top rookie bug: calling a reactive without parentheses (filtered) returns the function itself, not the value. Always filtered().

Day 4

Interactive Visualisations

library(plotly); library(DT)

ui <- fluidPage(plotlyOutput("scatter"), DTOutput("tbl"))

server <- function(input, output, session) {
  output$scatter <- renderPlotly({
    plot_ly(mtcars, x = ~wt, y = ~mpg, type = "scatter",
            mode = "markers", color = ~factor(cyl))
  })
  output$tbl <- renderDT({
    datatable(mtcars, filter = "top",
              options = list(pageLength = 10))
  })
}
                    

Day 5

shinydashboard and bslib

Professional UI without writing CSS

library(shinydashboard)

ui <- dashboardPage(
  dashboardHeader(title = "Sales Cockpit"),
  dashboardSidebar(
    sidebarMenu(
      menuItem("Overview",  tabName = "overview",  icon = icon("home")),
      menuItem("Customers", tabName = "customers", icon = icon("user")),
      menuItem("Products",  tabName = "products",  icon = icon("box"))
    )
  ),
  dashboardBody(
    tabItems(
      tabItem("overview",
        fluidRow(valueBoxOutput("rev"), valueBoxOutput("orders"), valueBoxOutput("aov")),
        fluidRow(box(plotOutput("trend"), width = 12))
      ),
      tabItem("customers", DTOutput("cust_tbl")),
      tabItem("products",  plotOutput("prod_chart"))
    )
  )
)
                    

For a more modern look, use bslib::bs_theme(version = 5, bootswatch = "minty") with bslib::cards() and layout_columns() - a polished Bootstrap-5 theme without writing any CSS.

Day 6

Deployment

From localhost to a public URL

shinyapps.io in five clicks

Create a free account at shinyapps.io.
In RStudio: Tools → Global Options → Publishing → Connect.
Open your app.R, click the blue Publish icon.
Pick the files to deploy. RStudio uploads and builds.
Open the URL it gives you. Done.

Before deploying: move large data into .rds files, pin packages with renv, wrap risky code in tryCatch(), add a showNotification() for slow operations.

Day 7

Project: Interactive Business Dashboard

Deploy a multi-tab Shiny dashboard for the e-commerce dataset:

Tab 1: KPI overview (revenue, AOV, orders, conversion) with date-range filter.
Tab 2: Region drill-down with cross-filtering.
Tab 3: Customer table (DT) with filters and CSV export.
Tab 4: Plotly scatter - units vs revenue, sized by margin.
Deployed to shinyapps.io with a public URL.

A full starter app lives at /data/week8-shiny-app-starter.R - download, adapt, ship.

Coming up: Week 9 - Power BI Fundamentals

The other BI tool you must know - and when to use it instead of R Shiny.

View Detailed Curriculum