Leaflet

R
intermediate
Published

July 6, 2026

No feedback found for this session

About this session

  • this is an intermediate introduction to Leaflet in Shiny as a tool for building interactive maps
  • you’ll need to be comfortable working in R Shiny to enjoy this session
TipSession resources

Why leaflet?

  • the easiest way of building interactive maps
  • mainly intended for point maps, but can also produce filled maps (choropleths)

Getting started

  1. Open a new Rstudio project
  2. In the console, run install.packages("pak")
  3. Once installation has completed, create a new temp.R script
  4. Install packages, and create some simple data using:
pak::pak(c("shiny", "leaflet", "dplyr", "tibble"))

library(tibble)
library(dplyr)
library(leaflet)

dat <- tibble::tibble(latitude = c(56, 57, 58),
                      long = c(-3, -4, -5),
                      val = c(1,2,3))

We’ll start with some standalone maps, and move on to thinking about how to add our maps to Shiny towards the end of the session.

A first map

Like ggplot, we build leaflet maps in layers. Start with a tibble of data, attach the main leaflet() function with a pipe, then add a second layer by piping on addMarkers() - beware the camelCase:

dat |>
  leaflet() |>
  addMarkers()

addMarkers() creates markers to correspond with our data. Usefully, it’ll interpret columns named lat(itude) and long(itude) as containing degrees of lat and long, so we’ve got a map of something here. But what’s not exactly clear.1 Let’s add an additional layer:

dat |>
  leaflet() |>
  addMarkers() |>
  addTiles()

It’s worth saying that, unlike ggplot, the order of layers in the code doesn’t determine the order of layers in the resulting visualisation - so your markers will still show up, even if we add tiles “over” them. That additional layer of addTiles() adds a background map for us, by default the OSM. That’s potentially change-able, using addProviderTiles():

dat |>
  leaflet() |>
  addMarkers() |>
  addProviderTiles(providers$NASAGIBS.ViirsEarthAtNight2012)

There’s a full list (and demo) of the different tiles at https://leaflet-extras.github.io/leaflet-providers/preview/index.html - those names get put after providers$.

Those map objects can be saved and recalled in the usual way:

base_map <- dat |>
  leaflet() |>
  addTiles() |>
  setView(lng = -5, lat = 57, zoom = 7) # lets you control starting location and zoom

base_map
base_map <- base_map |>
  addMarkers(icon = list(iconUrl = "src/images/icon.png", iconWidth = 25, iconHeight = 25)) # adding markers with a custom icon

base_map # re-view updated map
base_map |>
  addProviderTiles(providers$NASAGIBS.ViirsEarthAtNight2012) # add to overwrite

The custom icon here is just a small .png file with a transparent background: Custom icon

A more substantial map

Using this data set, we can build a map of Scotland’s hospitals:

hosp_dat <- here::here("r_training/data/locs.rds") |>
  readRDS() 

hosp_base <- hosp_dat |>
  leaflet() |>
  addTiles() |>
  addMarkers()

hosp_base

Tooltips

Our data contains hospital names, as well as long and lat values:

hosp_dat[1:5,] |>
  knitr::kable()
loc_name loc postcode long lat
University Hospital Ayr A210H KA6 6DX -4.595538 55.43033
University Hospital Crosshouse A111H KA2 0BE -4.539389 55.61394
Arran War Memorial Hospital A101H KA278LF -5.115493 55.54312
Davidson Cottage Hospital A207H KA269DS -4.851071 55.24244
Girvan Community Hospital A216H KA269HQ -4.846975 55.24831

Add a hospital name to the marker using a ~ to refer to the column name:

hosp_base |>
  addMarkers(label = ~loc_name)

If our data had unfriendly lat and long columns, we could also specify those in a similar way:

hosp_dat |>
  dplyr::rename(ad = long, ws = lat) |>
  leaflet() |>
  addTiles() |>
  addMarkers(label = ~loc_name, 
             lng = ~ad,
             lat = ~ws)

Unlike ggplot, we don’t have the full data to play with in leaflet objects, so we’ve needed to remake this map from scratch.

Customising markers

We often want to indicate group membership using marker colours. We’ll start by joining some board info to the hospitals data:

hosp_dat_board <- hosp_dat |>
  dplyr::left_join(readr::read_csv(here::here("r_training/data/board_locs.csv"))) |>
  dplyr::filter(!is.na(board)) # there are a few missing values

Next, and a major pain-point if you’re used to dealing with ggplot doing it for you, join a tibble of hex colour values:

board_colours <- tibble::tibble(board = unique(hosp_dat_board$board), 
               colours = c("#E6DFC8", "#D4A373", "#CCD5AE", "#E9EDC9", "#FAEDCD", "#B7B7A4", "#A8DADC", "#457B9D", "#E8AEB7", "#B388EB", "#1D3557", "#E63946", "#6B705C", "#3F37C9")
)

Then, build a similar map again, using a variant of addMarkers():

hosp_dat_board |>
  dplyr::left_join(board_colours) |>
  leaflet() |>
  addTiles() |>
  addCircleMarkers(label = ~loc_name, 
             stroke = T, # turns outline on and off
             color = "#000", # outline colour
             weight = 2, # outline thickness
             fillColor = ~colours, # column to use for fill colours
             fillOpacity = 0.8) # fill opacity

It’s also possible to scale those CircleMarkers. This uses some open data about hospital beds. This data is a bit fragmented, so we’ll remove the missing data and only show scaled blobs for about 30 sites in total here:

beds <- here::here("r_training/data/beds_site.csv") |>
  readr::read_csv() |>
  dplyr::filter(Quarter == "2025Q4") |>
  dplyr::select(loc = Location, SpecialtyName, beds = AverageAvailableStaffedBeds) |>
  dplyr::filter(!stringr::str_detect(loc, "S0")) |>
  dplyr::filter(SpecialtyName == "All Specialties") 

hosp_dat_boards_beds <- hosp_dat_board |>
  dplyr::left_join(board_colours) |>
  dplyr::left_join(beds) |>
  dplyr::filter(!is.na(beds)) 

hosp_dat_boards_beds |>
  leaflet() |>
  addTiles() |>
  addCircleMarkers(label = ~loc_name, 
             stroke = T, # turns outline on and off
             color = "#000", # outline colour
             weight = 2, # outline thickness
             fillColor = ~colours, # column to use for fill colours
             fillOpacity = 0.8, 
             radius = ~sqrt(beds))

Again, a reminder that you can change the base tiles to change the mapping used. Here’s an example of topographical mapping, showing very dramatically that people tend not to build big hospitals up mountains:

hosp_dat_boards_beds |>
  leaflet() |>
  addCircleMarkers(label = ~loc_name, 
             stroke = T, # turns outline on and off
             color = "#000", # outline colour
             weight = 2, # outline thickness
             fillColor = ~colours, # column to use for fill colours
             fillOpacity = 0.8, 
             radius = ~sqrt(beds)) |>
  addProviderTiles(providers$OpenTopoMap)

There’s also addAwesomeMarkers which uses fontawesome icons. Define the icon first using awesomeIcons, and then apply:

icons <- awesomeIcons(
  icon = 'building',
  library = 'fa',
  markerColor = "lightred",
  iconColor = "#ffffff")

hosp_dat_board |>
  dplyr::left_join(board_colours) |>
  leaflet() |>
  addTiles() |>
  addAwesomeMarkers(label = ~loc_name, 
             icon = icons) # using our earlier definition

addAwesomeMarkers is hard to colour with precision (or to fit series), as you can only use a group of a dozen or so basic colours for the markerColor aspect. That’s also a good time to demonstrate clustering:

hosp_dat_board |>
  dplyr::left_join(board_colours) |>
  leaflet() |>
  addTiles() |>
  addAwesomeMarkers(label = ~loc_name, 
             group = ~board, 
             clusterOptions = markerClusterOptions())

Adding leaflet to Shiny

  • data source at https://nes-dew.github.io/KIND-training/r_training/data/hosp_dat_boards_beds.rds
  • leafletOutput and renderLeaflet to output and render your leaflet map
  • using sidebarLayout with sidebarPanel and mainPanel is highly recommended - leaflet otherwise defaults to a full-screen view
    • or bslib has page_sidebar which is even nicer
  • setView is helpful if you want to constrain the zoom level of your map while filtering

Footnotes

  1. If, like me, you struggle to remember which one is which, longitude is the East-West one. Earth is slightly wider than it is tall - an oblate spheroid, famously - so the longitude is slightly longer than the other one. Lots of interesting history about longitude too - it’s probably the most-studied bit of history of science of all.↩︎