tidyterra

The goal of {tidyterra} is to provide common methods of the tidyverse packages for objects created with the {terra} package: SpatRaster and SpatVector. It also provides geoms for plotting these objects with {ggplot2}.

Overview

You can have a look to the documentation of the dev version in https://dieghernan.github.io/tidyterra/dev/

{tidyverse} methods implemented on {tidyterra} works differently depending on the type of Spat* object:

• SpatVector: the methods are implemented using terra::as.data.frame() coercion. Rows correspond to geometries and columns correspond to attributes of the geometry.

• SpatRaster: The implementation on SpatRaster objects differs, since the methods could be applied to layers or to cells. {tidyterra} overall approach is to treat the layers as columns of a tibble and the cells as rows (i.e. select(SpatRaster, 1) would select the first layer of a SpatRaster).

The methods implemented return the same type of object used as input, unless the expected behavior of the method is to return another type of object, (for example, as_tibble() would return a tibble).

Current methods and functions provided by {tidyterra} are:

❗ A note on performance

{tidyterra} is conceived as a user-friendly wrapper of {terra} using the {tidyverse} methods and verbs. This approach therefore has a cost in terms of performance.

If you are a heavy user of {terra} or you need to work with big raster files, {terra} is much more focused on terms of performance. When possible, each function of {tidyterra} references to its equivalent on {terra}.

As a rule of thumb if your raster has less than 10.000.000 data slots counting cells and layers (i.e. terra::ncell(your_rast)*terra::nlyr(your_rast) < 10e6) you are good to go with {tidyterra}.

When plotting rasters, resampling is performed automatically (as terra::plot() does, see the help page). You can adjust this with the maxcell parameter.

Installation

Install {tidyterra} from CRAN:

install.packages("tidyterra")

You can install the development version of {tidyterra} like so:

remotes::install_github("dieghernan/tidyterra")

Alternatively, you can install {tidyterra} using the r-universe:

# Enable this universe
options(repos = c(
  dieghernan = "https://dieghernan.r-universe.dev",
  CRAN = "https://cloud.r-project.org"
))
install.packages("tidyterra")

Example

SpatRasters

This is a basic example which shows you how to manipulate and plot SpatRaster objects:

library(tidyterra)

library(terra)


# Temperatures
f <- system.file("extdata/cyl_temp.tif", package = "tidyterra")

rastertemp <- rast(f)

library(ggplot2)

# Facet all layers

ggplot() +
  geom_spatraster(data = rastertemp) +
  facet_wrap(~lyr, ncol = 2) +
  scale_fill_whitebox_c(
    palette = "muted",
    labels = scales::label_number(suffix = "º")
  ) +
  labs(fill = "Avg temperature")

# Create maximum differences

variation <- rastertemp %>%
  mutate(
    diff = tavg_06 - tavg_04
  ) %>%
  select(var_apr_jun = diff)

# Add also a overlay of a SpatVector
f_vect <- system.file("extdata/cyl.gpkg", package = "tidyterra")

prov <- vect(f_vect)

ggplot(prov) +
  geom_spatraster(data = variation) +
  geom_spatvector(fill = NA) +
  scale_fill_whitebox_c(
    palette = "deep", direction = -1,
    labels = scales::label_number(suffix = "º")
  ) +
  theme_minimal() +
  coord_sf(crs = 25830) +
  labs(
    fill = "Difference",
    title = "Variation of temperature in Castille and Leon (Spain)",
    subtitle = "(Average) temperatures in June vs. April"
  )

{tidyterra} also provide a geom for plotting RGB SpatRaster tiles with {ggplot2}

f_tile <- system.file("extdata/cyl_tile.tif", package = "tidyterra")

rgb_tile <- rast(f_tile)


plot <- ggplot(prov) +
  geom_spatraster_rgb(data = rgb_tile) +
  geom_spatvector(fill = NA) +
  theme_light()


plot

# Recognizes coord_sf()
plot +
  # Change crs and datum (for relabeling graticules)
  coord_sf(crs = 3857, datum = 3857)

{tidyterra} provides specific scales for plotting hypsometric maps with {ggplot2}:

asia <- rast(system.file("extdata/asia.tif", package = "tidyterra"))

terra::plot(asia)

ggplot() +
  geom_spatraster(data = asia) +
  scale_fill_hypso_tint_c(
    palette = "gmt_globe",
    labels = scales::label_number(),
    breaks = c(-10000, -5000, 0, 2500, 5000, 8000),
    guide = guide_colorbar(
      direction = "horizontal",
      title.position = "top",
      barwidth = 20
    )
  ) +
  labs(
    fill = "elevation (m)",
    title = "Hypsometric map of Asia"
  ) +
  theme_minimal() +
  theme(legend.position = "bottom")

SpatVectors

This is a basic example which shows you how to manipulate and plot SpatVector objects:

vect(system.file("ex/lux.shp", package = "terra")) %>%
  group_by(NAME_1) %>%
  summarise(pop_dens = sum(POP) / sum(AREA)) %>%
  glimpse() %>%
  autoplot(aes(fill = pop_dens)) +
  scale_fill_whitebox_c(palette = "pi_y_g")
#> Geometry type: Polygons
#> Geodetic CRS: lon/lat WGS 84 (EPSG:4326)
#> Extent (x , y) : [5° 44' 38.9045" E - 6° 31' 41.7076" E] , [49° 26' 52.1063" N - 50° 10' 53.8376" N]
#> Rows: 3
#> Columns: 2
#> $ NAME_1   <chr> "Diekirch", "Grevenmacher", "Luxembourg"
#> $ pop_dens <dbl> 80.83865, 134.90133, 485.34879

I need your feedback

{tidyterra} is currently on development mode. Please leave your feedback or open an issue on https://github.com/dieghernan/tidyterra/issues.

Need help?

Check our FAQs or open a new issue!

Citation

To cite ‘tidyterra’ in publications use:

Hernangomez D (2023). tidyterra: tidyverse Methods and ggplot2 Helpers for terra Objects. https://doi.org/10.5281/zenodo.6572471, https://dieghernan.github.io/tidyterra/

A BibTeX entry for LaTeX users is

@Manual{R-tidyterra,
  title = {{tidyterra}: tidyverse Methods and ggplot2 Helpers for terra Objects},
  doi = {10.5281/zenodo.6572471},
  author = {Diego Hernangómez},
  year = {2023},
  version = {0.4.0.9000},
  url = {https://dieghernan.github.io/tidyterra/},
  abstract = {Extension of the tidyverse for SpatRaster and SpatVector objects of the terra package. It includes also new geom_ functions that provide a convenient way of visualizing terra objects with ggplot2.},
}

Acknowledgement

{tidyterra} ggplot2 geoms are based on {ggspatial} implementation, by Dewey Dunnington and ggspatial contributors.