Awesome

tidyterra <a href="https://dieghernan.github.io/tidyterra/"><img src="man/figures/logo.png" align="right" height="139"/></a>

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.

Please cite tidyterra as:

Hernangómez, D., (2023). Using the tidyverse with terra objects: the tidyterra package. Journal of Open Source Software, 8(91), 5751, https://doi.org/10.21105/joss.05751.

A BibTeX entry for LaTeX users is:

@article{Hernangómez2023,
  doi = {10.21105/joss.05751},
  url = {https://doi.org/10.21105/joss.05751},
  year = {2023},
  publisher = {The Open Journal},
  volume = {8},
  number = {91},
  pages = {5751},
  author = {Diego Hernangómez},
  title = {Using the {tidyverse} with {terra} objects: the {tidyterra} package},
  journal = {Journal of Open Source Software}
}

Overview

Full manual of the most recent release of tidyterra on CRAN is online: https://dieghernan.github.io/tidyterra/

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:

:exclamation: 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
install.packages("tidyterra", repos = c(
  "https://dieghernan.r-universe.dev",
  "https://cloud.r-project.org"
))

Example

SpatRasters

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

library(tidyterra)
library(terra)

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

rastertemp
#> class       : SpatRaster 
#> dimensions  : 87, 118, 3  (nrow, ncol, nlyr)
#> resolution  : 3881.255, 3881.255  (x, y)
#> extent      : -612335.4, -154347.3, 4283018, 4620687  (xmin, xmax, ymin, ymax)
#> coord. ref. : World_Robinson 
#> source      : cyl_temp.tif 
#> names       :   tavg_04,   tavg_05,  tavg_06 
#> min values  :  1.885463,  5.817587, 10.46338 
#> max values  : 13.283829, 16.740898, 21.11378

# Rename
rastertemp <- rastertemp %>%
  rename(April = tavg_04, May = tavg_05, June = tavg_06)

# Facet all layers
library(ggplot2)

ggplot() +
  geom_spatraster(data = rastertemp) +
  facet_wrap(~lyr, ncol = 2) +
  scale_fill_whitebox_c(
    palette = "muted",
    labels = scales::label_number(suffix = "º"),
    n.breaks = 12,
    guide = guide_legend(reverse = TRUE)
  ) +
  labs(
    fill = "",
    title = "Average temperature in Castille and Leon (Spain)",
    subtitle = "Months of April, May and June"
  )

# Create maximum differences of two months
variation <- rastertemp %>%
  mutate(diff = June - May) %>%
  select(variation = diff)

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

ggplot(prov) +
  geom_spatraster(data = variation) +
  geom_spatvector(fill = NA) +
  scale_fill_whitebox_c(
    palette = "deep", direction = -1,
    labels = scales::label_number(suffix = "º"),
    n.breaks = 5
  ) +
  theme_minimal() +
  coord_sf(crs = 25830) +
  labs(
    fill = "variation",
    title = "Variation of temperature in Castille and Leon (Spain)",
    subtitle = "Average temperatures in June vs. May"
  )

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

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

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(),
    # Further refinements
    breaks = c(-10000, -5000, 0, 2000, 5000, 8000),
    guide = guide_colorbar(reverse = TRUE)
  ) +
  labs(
    fill = "elevation (m)",
    title = "Hypsometric map of Asia"
  ) +
  theme(
    legend.position = "bottom",
    legend.title.position = "top",
    legend.key.width = rel(3),
    legend.ticks = element_line(colour = "black", linewidth = 0.3),
    legend.direction = "horizontal"
  )

SpatVectors

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

vect(system.file("ex/lux.shp", package = "terra")) %>%
  mutate(pop_dens = POP / AREA) %>%
  glimpse() %>%
  autoplot(aes(fill = pop_dens)) +
  scale_fill_whitebox_c(palette = "pi_y_g") +
  labs(
    fill = "population per km2",
    title = "Population density of Luxembourg",
    subtitle = "By canton"
  )
#> #  A SpatVector 12 x 7
#> #  Geometry type: Polygons
#> #  Geodetic CRS: lon/lat WGS 84 (EPSG:4326)
#> #  Extent (x / y) : ([5° 44' 38.9" E / 6° 31' 41.71" E] , [49° 26' 52.11" N / 50° 10' 53.84" N])
#> 
#> $ ID_1     <dbl> 1, 1, 1, 1, 1, 2, 2, 2, 3, 3, 3, 3
#> $ NAME_1   <chr> "Diekirch", "Diekirch", "Diekirch", "Diekirch", "Diekirch", "…
#> $ ID_2     <dbl> 1, 2, 3, 4, 5, 6, 7, 12, 8, 9, 10, 11
#> $ NAME_2   <chr> "Clervaux", "Diekirch", "Redange", "Vianden", "Wiltz", "Echte…
#> $ AREA     <dbl> 312, 218, 259, 76, 263, 188, 129, 210, 185, 251, 237, 233
#> $ POP      <int> 18081, 32543, 18664, 5163, 16735, 18899, 22366, 29828, 48187,…
#> $ pop_dens <dbl> 57.95192, 149.27982, 72.06178, 67.93421, 63.63118, 100.52660,…

I need your feedback

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!

You can also ask in Stack Overflow using the tag [tidyterra].

Acknowledgement

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