Get started with leaflet.indoor • leaflet.indoor

leaflet.indoor adds indoor vector layers to a leaflet map and keeps one floor visible per indoor data set. It works with saved HTML widgets, R Markdown, Quarto, and Shiny without a tile service.

Data structure

The primary input is an sf object. A column named level is used by default. Numeric, character, and factor floors are supported. A list-column assigns one feature to more than one floor:

indoor_demo[c("name", "level")]
#> Simple feature collection with 10 features and 2 fields
#> Geometry type: GEOMETRY
#> Dimension:     XY
#> Bounding box:  xmin: 0 ymin: 0 xmax: 17 ymax: 23
#> CRS:           NA
#>                     name   level                       geometry
#> 1           Meeting room       0 POLYGON ((0 0, 8 0, 8 6, 0 ...
#> 2             Laboratory       0 POLYGON ((9 0, 17 0, 17 6, ...
#> 3          Main corridor       0         LINESTRING (0 7, 17 7)
#> 4              Stairwell 0, 1, 2 POLYGON ((7.5 6, 9.5 6, 9.5...
#> 5               Elevator 0, 1, 2                   POINT (12 8)
#> 6           Reading room       1 POLYGON ((0 10, 8 10, 8 16,...
#> 7               Workshop       1 POLYGON ((9 10, 17 10, 17 1...
#> 8          Service suite       2 MULTIPOLYGON (((0 18, 6 18,...
#> 9        Emergency route       2 MULTILINESTRING ((0 17, 17 ...
#> 10 Environmental sensors       0     MULTIPOINT ((3 3), (14 3))

The included data are synthetic. Rooms use polygons, the corridor and emergency route use line geometries, the elevator uses a point, and the sensor feature is a multipoint. The stairwell and elevator belong to floors 0, 1, and 2.

Missing or empty floors are errors by default. Use missing_level = "drop" only when dropping incomplete features is intentional.

A local Cartesian map

Local plans must use L.CRS.Simple, have no declared sf CRS, and explicitly set crs = "simple". Coordinates are passed through unchanged.

leaflet(
  indoor_demo,
  options = leafletOptions(
    crs = leafletCRS(crsClass = "L.CRS.Simple"),
    minZoom = -2
  )
) |>
  addIndoor(
    level = "level",
    initial_level = "0",
    layerId = ~feature_id,
    label = ~name,
    popup = ~description,
    style = list(fillColor = ~fill),
    crs = "simple"
  ) |>
  addIndoorControl(position = "bottomright")

Floor order

Pure numeric levels are ordered numerically. Character levels use a stable natural order, including basements such as B2 and B1. Factor levels retain their observed factor order. For domain-specific labels, provide every observed floor exactly once:

addIndoor(
  level_order = c("Basement", "Ground", "Mezzanine", "Upper"),
  initial_level = "Ground"
)

The canonical order runs from lowest to highest. The vertical control displays the reverse order by default so higher floors appear above lower floors. Set indoorControlOptions(descending = FALSE) to display canonical order.

Styles, labels, and popups

indoorOptions() defines separate defaults for points, lines, and polygons. The style list can contain constants, vectors, or formulas. Its values override the geometry defaults feature by feature.

Plain labels and popups are escaped. Mark up trusted content explicitly:

addIndoor(
  label = ~name,
  popup = htmltools::HTML("<strong>Trusted content</strong>"),
  style = list(fillColor = ~fill, weight = 2)
)

For paired room images and captions, see the vignette("photo-carousels", package = "leaflet.indoor") article.

Multiple indoor data sets

Data sets are independent when they use different dataset_id values. Controls are independent when they use different control_id values. Identifiers are scoped to one map, so separate widgets never share an active floor.

map |>
  addIndoor(data = rooms, dataset_id = "rooms") |>
  addIndoorControl(dataset_id = "rooms", control_id = "rooms-control") |>
  addIndoor(data = services, dataset_id = "services") |>
  addIndoorControl(dataset_id = "services", control_id = "services-control")

Geographic data

sf objects with a declared geographic or projected CRS are transformed to EPSG:4326 for an ordinary Leaflet map. This example intentionally omits online tiles so it builds without network access.

leaflet(indoor_demo_geo) |>
  addIndoor(
    layerId = ~feature_id,
    label = ~name,
    popup = ~description,
    style = list(fillColor = ~fill)
  ) |>
  addIndoorControl()

Shiny and proxies

Use setIndoorLevel() with a proxy after a map has rendered:

leafletProxy("map") |>
  setIndoorLevel("2")

User changes arrive in input$map_indoor_level as a list containing level, dataset_id, and control_id. See vignette("shiny", package = "leaflet.indoor") for a complete application.

Troubleshooting

A missing level column is an error and names the missing column.
A local plan on a geographic map is rejected instead of being silently interpreted as longitude and latitude.
A spatial CRS on an L.CRS.Simple map is rejected because no defensible spatial transformation exists.
An unknown floor passed to a regular map is an R error. On a Shiny proxy it is rejected in the browser and leaves the current floor unchanged.

Known limitations and roadmap

Version 0.1.0 supports two-dimensional GeoJSON-compatible vector geometry. It does not provide rasters, three-dimensional floors, indoor routing, feature editing, or automatic alignment of unrelated floor plans. Future work will be driven by reproducible use cases and compatibility with the Leaflet JavaScript version embedded by the R leaflet package.