Skip to contents

Get started

Source: vignettes/articles/aobview.Rmd
aobview.Rmd

aobview is in early development. This page shows what works today; the home page lists what is still being worked out.

Install

aobview and its core package, aobcore, are installed from GitHub:

# install.packages("remotes")
remotes::install_github("allboa/aobview")

sf is needed for vector data, and terra and gdalraster for rasters.

A first view

view() takes an sf object and draws it in a page. aobcore ships a coastline south of 40S in longitude and latitude; aobview sees that it lies entirely in the far south and draws it in Antarctic Polar Stereographic (EPSG:3031), densifying edges first so that parallels curve.

library(aobview)
coast <- sf::st_read(
  system.file("extdata", "coastline_south_40s.geojson", package = "aobcore"),
  quiet = TRUE
)
v <- view(coast)

In an interactive session, printing the view opens the page in the IDE’s viewer or the browser. v$file is the page and v$scene the scene it draws. The page needs no server, and no network unless it references a remote COG.

Colour, legends and popups

zcol colours features by an attribute and adds a legend; clicking or tapping a feature shows its attributes. Here are the CCAMLR statistical areas (a small, simplified copy shipped with aobview, in EPSG:6932), coloured by area and drawn in EPSG:3031 with the coastline on top.

areas <- sf::st_read(
  system.file("extdata", "ccamlr_statistical_areas.geojson", package = "aobview"),
  quiet = TRUE
)
areas$area <- paste0("Area ", substr(areas$GAR_Long_Label, 1, 2))
v <- view(areas, crs = "EPSG:3031", zcol = "area",
          popup = c("GAR_Name", "GAR_Long_Label", "GAR_Start_Date", "GAR_Size")) |>
  view_add(coast, popup = FALSE)

Numeric columns take a continuous palette, or classes with breaks =; palette is any grDevices palette name.

v <- view(areas, zcol = "GAR_Size", palette = "Viridis")

That view is in the data’s own CRS, EPSG:6932, because a projected CRS is kept as it is.

Any CRS

crs = draws the same data in any projected CRS, given as an authority code, a PROJ string or WKT. A Lambert azimuthal equal-area view centred on the pole, with 140E at the top:

v <- view(coast, crs = "+proj=laea +lat_0=-90 +lon_0=140 +datum=WGS84")

Rasters

A terra raster is planned into tiles by aobcore and drawn on meshes in the view CRS, so it is never resampled in R. A remote Cloud Optimized GeoTIFF is referenced by URL and the browser fetches its tiles; a local or in-memory raster is embedded. This one is a small polar field in EPSG:3031 shipped with aobcore, with the coastline over it: the v1 target of a polar raster with vector overlays, in one call.

sst <- terra::rast(system.file("extdata", "polar_3031.tif", package = "aobcore"))
v <- view(list(sst = sst, coastline = coast))

Raster views need gdalraster with a working PROJ database. On macOS, CRAN’s gdalraster binary cannot currently find its proj.db; gdalraster from conda-forge works.

Where next