HTML page layouts

This section will discuss how to customize the overall layout and appearance of shiny apps.

Common structure

The code below show a common setup for shiny apps—a page with a sidebar and main panel— along with a graph of how the pieces get laid out on a webpage.

app_ui = ui.page_fluid(
    ui.panel_title(),
    ui.layout_sidebar(
        ui.panel_sidebar(
            ...
        ),
        ui.panel_main(
            ...
        ),
    ),
)
page_fluid()
page_fluid()
  panel_title()
  panel_title()
layout_sidebar()
layout_…
panel_sidebar()
panel_s…
panel_main()
panel_m…
Text is not SVG - cannot display

Notice that the first piece of each function name gives a hint about what it does:

  • page_* is the outermost piece. page_fluid means it will expand to fill the full width of the browser window, rather than stopping at a certain width.
  • layout_* positions the pieces inside it (e.g. put them side-by-side).
  • panel_* is used for a range of common pieces used in shiny apps.

Shiny applications often use page_fluid(), so we’ll focuses on other aspects of laying out pages, before discussing choosing between page_* functions at the end.

Page with sidebar

A common approach for shiny applications is to put inputs in a sidebar, and then output content on the main section of the page.

This is shown in the application below.

#| standalone: true
#| components: [editor, viewer]
#| layout: vertical
#| viewerHeight: 600

from shiny import App, render, ui
import matplotlib.pyplot as plt
import numpy as np

app_ui = ui.page_fluid(
    ui.panel_title("Simulate a normal distribution"),

    ui.layout_sidebar(

      ui.panel_sidebar(
        ui.input_slider("n", "Sample size", 0, 1000, 250),
        ui.input_numeric("mean", "Mean", 0),
        ui.input_numeric("std_dev", "Standard deviation", 1),
        ui.input_slider("n_bins", "Number of bins", 0, 100, 20),
      ),

      ui.panel_main(
        ui.output_plot("plot")
      ),
    ),
)


def server(input, output, session):

    @output
    @render.plot(alt="A histogram")
    def plot() -> object:
        x = np.random.normal(input.mean(), input.std_dev(), input.n())

        fig, ax = plt.subplots()
        ax.hist(x, input.n_bins(), density=True)
        return fig


app = App(app_ui, server)

Note that if the browser window is narrow, the sidebar will be shown on top of the main panel.

Adding rows and columns

In addition to defining big pieces—like a sidebar—it’s quick to add rows and columns to your content. You can set pieces side-by-side, while controlling the width of each piece, while ensuring it adapts well to narrow screens (e.g. mobile devices).

This is shown in the app below, which has two rows and columns of different widths.

#| standalone: true
#| components: [editor, viewer]
#| layout: vertical
#| viewerHeight: 200

from shiny import App, ui
import matplotlib.pyplot as plt
import numpy as np

style="border: 1px solid #999;"

app_ui = ui.page_fluid(
    ui.row(
        ui.column(4, "row-1 col-1", style=style),
        ui.column(8, "row-1 col-2", style=style),
    ),
    ui.row(
        ui.column(6, "row-2 col-1", style=style),
        ui.column(6, "row-2 col-2", style=style),
    ),
)


app = App(app_ui, None)

Notice that we first define rows, then columns inside them. Importantly, the first parameter to ui.column() is how wide (relative to others in the row) each column should be. Rows are defined as 12 units wide.

Choosing a page function

While page_fluid() is the most common page function used, below is a description of the different page function options.

name description
page_fluid() Continuously expand width to fit the screen.
page_fixed() Expand width at certain pre-defined breakpoints. (E.g. when shifting from “small” to “medium” width.)
page_nav() Create a page with a top navigation bar. Note that navigation is discussed in detail in the next section.
page_bootstrap() A basic page that makes sure bootstrap is loaded. Customization is left to the user.