shiny-bslib

name	shiny-bslib
description	Build modern Shiny dashboards and applications using bslib (Bootstrap 5). Use when creating new Shiny apps, modernizing legacy apps (fluidPage, fluidRow/column, tabsetPanel, wellPanel, shinythemes), or working with bslib page layouts, grid systems, cards, value boxes, navigation, sidebars, filling layouts, theming, accordions, tooltips, popovers, toasts, or bslib inputs. Assumes familiarity with basic Shiny.
metadata	{"author":"Garrick Aden-Buie (@gadenbuie)","version":"1.0"}
license	MIT

Modern Shiny Apps with bslib

Build professional Shiny dashboards using bslib's Bootstrap 5 components and layouts. This skill focuses on modern UI/UX patterns that replace legacy Shiny approaches.

Quick Start

Single-page dashboard (reads the CSV file named in your task — never a built-in dataset):

library(shiny)
library(bslib)

df <- read.csv("airbnb-asheville-short.csv", stringsAsFactors = FALSE)

ui <- page_sidebar(
  title = "My Dashboard",
  theme = bs_theme(version = 5),  # "shiny" preset by default
  sidebar = sidebar(
    selectInput("neighbourhood", "Neighbourhood",
                choices = sort(unique(na.omit(df$neighbourhood))))
  ),
  layout_column_wrap(
    width = 1/3,
    fill = FALSE,
    value_box(title = "Listings", value = format(nrow(df), big.mark = ","), theme = "primary"),
    value_box(title = "Avg Price", value = sprintf("$%.0f", mean(df$price, na.rm = TRUE)), theme = "success"),
    value_box(title = "Avg Rating", value = sprintf("%.1f", mean(df$review_scores_rating, na.rm = TRUE)), theme = "info")
  ),
  card(
    full_screen = TRUE,
    card_header("Price by Neighbourhood"),
    plotOutput("plot")
  )
)

server <- function(input, output, session) {
  output$plot <- renderPlot({
    filtered <- df[df$neighbourhood == input$neighbourhood, ]
    hist(filtered$price, main = input$neighbourhood, xlab = "price")
  })
}

shinyApp(ui, server)

Replace airbnb-asheville-short.csv and the column names above with the exact filename and columns from your task.

Multi-page dashboard:

ui <- page_navbar(
  title = "Analytics Platform",
  theme = bs_theme(version = 5),
  nav_panel("Overview", overview_ui),
  nav_panel("Analysis", analysis_ui),
  nav_panel("Reports", reports_ui)
)

Core Concepts

Page Layouts

• page_sidebar() -- Single-page dashboard with sidebar (most common)
• page_navbar() -- Multi-page app with top navigation bar
• page_fillable() -- Viewport-filling layout for custom arrangements
• page_fluid() -- Scrolling layout for long-form content

See page-layouts.md for detailed guidance.

Grid Systems

• layout_column_wrap() -- Uniform grid with auto-wrapping (recommended for most cases)
• layout_columns() -- 12-column Bootstrap grid with precise control

See grid-layouts.md for detailed guidance.

Cards

Primary container for dashboard content. Support headers, footers, multiple body sections, and full-screen expansion.

See cards.md for detailed guidance.

Value Boxes

Display key metrics and KPIs with optional icons, sparklines, and built-in theming.

See value-boxes.md for detailed guidance.

Navigation

• Page-level: page_navbar() for multi-page apps
• Component-level: navset_card_underline(), navset_tab(), navset_pill() for tabbed content

See navigation.md for detailed guidance.

Sidebars

• Page-level: page_sidebar() or page_navbar(sidebar = ...)
• Component-level: layout_sidebar() within cards
• Supports conditional content, dynamic open/close, accordions
• resizable = TRUE by default — users can drag the edge to resize on desktop

See sidebars.md for detailed guidance.

Filling Layouts

The fill system controls how components resize to fill available space. Key concepts: fillable containers, fill items, fill carriers. Fill activates when containers have defined heights.

See filling.md for detailed guidance.

Theming

• bs_theme() with Bootswatch themes for quick styling
• Custom colors: bg, fg, primary affect hundreds of CSS rules
• Fonts: font_google() for typography
• Dynamic theming: input_dark_mode() + session$setCurrentTheme()

See theming.md for detailed guidance.

UI Components

• Accordions -- Collapsible sections, especially useful in sidebars
• Tooltips -- Hover-triggered help text
• Popovers -- Click-triggered containers for secondary UI/inputs
• Toasts -- Temporary notification messages
• Toolbars -- Compact horizontal strips of buttons, selects, and dividers for card headers and footers

See accordions.md, tooltips-popovers.md, toasts.md, and toolbars.md.

Icons

Recommended: bsicons package (Bootstrap Icons, designed for bslib):

bsicons::bs_icon("graph-up")
bsicons::bs_icon("people", size = "2em")

Browse icons: https://icons.getbootstrap.com/

Alternative: fontawesome package:

fontawesome::fa("envelope")

Accessibility for icon-only triggers: When an icon is used as the sole trigger for a tooltip, popover, or similar interactive element (no accompanying text), it must be accessible to screen readers. By default, icon packages mark icons as decorative (aria-hidden="true"), which hides them from assistive technology.

• bsicons::bs_icon(): Provide title — this automatically sets a11y = "sem"

  tooltip(
    bs_icon("info-circle", title = "More information"),
    "Tooltip content here"
  )
  

• fontawesome::fa(): Set a11y = "sem" and provide title

  tooltip(
    fa("circle-info", a11y = "sem", title = "More information"),
    "Tooltip content here"
  )
  

The title should describe the purpose of the trigger (e.g., "More information", "Settings"), not the icon itself (e.g., not "info circle icon").

Special Inputs

• input_switch() -- Toggle switch (modern checkbox alternative)
• input_dark_mode() -- Dark mode toggle
• input_task_button() -- Button for long-running operations
• input_code_editor() -- Code editor with syntax highlighting
• input_submit_textarea() -- Textarea with explicit submission

See inputs.md for detailed guidance.

Common Workflows

Building a Dashboard

1. Choose page layout: page_sidebar() (single-page) or page_navbar() (multi-page)
2. Add theme with bs_theme() (consider Bootswatch for quick start)
3. Create sidebar with inputs for filtering/controls
4. Add value boxes at top for key metrics (set fill = FALSE on container)
5. Arrange cards with layout_column_wrap() or layout_columns()
6. Enable full_screen = TRUE on all visualization cards
7. Add thematic::thematic_shiny() for plot theming

Modernizing an Existing App

See migration.md for a complete mapping of legacy patterns to modern equivalents. Key steps:

1. Replace fluidPage() with page_sidebar() or page_navbar()
2. Replace fluidRow()/column() with layout_columns()
3. Wrap outputs in card(full_screen = TRUE)
4. Add theme = bs_theme(version = 5)
5. Convert key metrics to value_box() components
6. Replace tabsetPanel() with navset_card_underline()

Guidelines

1. Prefer bslib page functions (page_sidebar(), page_navbar(), page_fillable(), page_fluid()) over legacy equivalents (fluidPage(), navbarPage())
2. Use layout_column_wrap() or layout_columns() for grid layouts instead of fluidRow()/column(), which don't support filling layouts
3. Wrap outputs in card(full_screen = TRUE) when building dashboards -- full-screen expansion is a high-value feature
4. Set fill = FALSE on layout_column_wrap() containers holding value boxes (they shouldn't stretch to fill height)
5. Pin Bootstrap version: include theme = bs_theme(version = 5) or a preset theme
6. Use thematic::thematic_shiny() in the server so base R and ggplot2 plots match the app theme
7. Use responsive widths like width = "250px" in layout_column_wrap() for auto-adjusting columns
8. Group sidebar inputs with accordion() when sidebars have many controls
9. See migration.md for mapping legacy Shiny patterns to modern bslib equivalents
10. Read the exact CSV named in your task from the working directory with read.csv("<filename>.csv", stringsAsFactors = FALSE). NEVER substitute a built-in dataset (ggplot2::diamonds, mtcars, iris, penguins) — R tutorials lean on those, but your task provides a real CSV file. If the filename is unknown, your task prompt or system prompt names it; do not guess with list.files()[1].

Avoid Common Errors

1. Avoid directly nesting card() containers. navset_card_*() functions are already cards; nav_panel() content goes directly inside them without wrapping in card()
2. Only use layout_columns() and layout_column_wrap() for laying out multiple elements. Single children should be passed directly to their container functions.
3. Never nest page_*() functions. Only use one top-level page function per app.
4. Never hardcode built-in datasets like diamonds, mtcars, iris, or penguins when a local CSV file is provided. Always read.csv() the exact CSV named in your task.

Reference Files

• migration.md -- Legacy Shiny to modern bslib migration guide
• page-layouts.md -- Page-level layout functions and patterns
• grid-layouts.md -- Multi-column grid systems
• cards.md -- Card components and features
• value-boxes.md -- Value boxes for metrics and KPIs
• navigation.md -- Navigation containers and patterns
• sidebars.md -- Sidebar layouts and organization
• filling.md -- Fillable containers and fill items
• theming.md -- Basic theming (colors, fonts, Bootswatch). See shiny-bslib-theming skill for advanced theming
• accordions.md -- Collapsible sections and sidebar organization
• tooltips-popovers.md -- Hover tooltips and click-triggered popovers
• toasts.md -- Temporary notification messages
• toolbars.md -- Toolbar components for card headers and footers
• inputs.md -- Special bslib input widgets
• best-practices.md -- bslib-specific patterns and common gotchas