Skip to content

Latest commit

 

History

History
224 lines (187 loc) · 7.06 KB

README.md

File metadata and controls

224 lines (187 loc) · 7.06 KB

vectormetrics

Lifecycle: experimental Codecov test coverage R-CMD-check

Overview

vectormetrics is an R package for calculating landscape and shape metrics for vector layers. Its aim is to provide a set of metrics that can characterize landscape patterns and properties of the shapes defined as polygons and multipolygons. Whole package is based on Simple Feature geometry standard provided by sf package. Every function can be used in a tidy, piped workflow, as it always takes the data as the first argument and returns a tibble.

Installation

You can download most recent development version of the package from GitHub with:

remotes::install_github("r-spatialecology/vectormetrics")

Using vectormetrics

Function names structure

All functions in vectormetrics start with vm_ (for vector metrics). The second part of the name specifies the level (patch - p, class - c or landscape - l). The last part of the function name is the abbreviation of the corresponding metric (e.g. enn for the euclidean nearest-neighbor distance and rect for the rectangularity). Some landscape and class level functions have also a suffix at the end, that specifies the aggregation method (e.g. mean, sd).

# Patch level
## vm_p_"metric"
vm_p_area()
vm_p_square()

# Class level
## vm_c_"metric"[_"aggregation"]
vm_c_np()
vm_c_shape_sd()

# Landscape level
## vm_l_"metric"[_"aggregation"]
vm_l_lpi()
vm_l_square_mn()

Examples

Some examples of calculating metrics on all levels and with different class and patch columns.

library(vectormetrics)
library(sf)
data("vector_landscape")
plot(vector_landscape)

## Shape index
vm_p_shape(vector_landscape, class_col = "class")
#> MULTIPOLYGON geometry provided. You may want to cast it to separate polygons with 'get_patches()'.
#> This message is displayed once per session.
#> # A tibble: 3 × 5
#>   level class id    metric value
#>   <chr> <chr> <chr> <chr>  <dbl>
#> 1 patch 1     1     shape   5.06
#> 2 patch 2     2     shape   4.76
#> 3 patch 3     3     shape   4.80

## Number of patches
vm_c_np(vector_landscape, class_col = "class")
#> # A tibble: 3 × 5
#>   level class id    metric value
#>   <chr> <chr> <chr> <chr>  <int>
#> 1 class 1     <NA>  np         1
#> 2 class 2     <NA>  np         1
#> 3 class 3     <NA>  np         1

## Largest patch index
vm_l_lpi(vector_landscape)
#> # A tibble: 1 × 5
#>   level     class id    metric value
#>   <chr>     <chr> <chr> <chr>  <dbl>
#> 1 landscape <NA>  <NA>  lpi     49.7

## Mean squareness
vm_l_square_mn(vector_landscape)
#> # A tibble: 1 × 5
#>   level     class id    metric value
#>   <chr>     <chr> <chr> <chr>  <dbl>
#> 1 landscape <NA>  <NA>  sq_mn  0.232

Utility functions

For now there are two utility functions available in the package. First one is get_patches() which breaks multipolygon geometries into polygons. There are two types of neighborhood relations available: 4 (edge) and 8 (vertex). This function enables users to create set of geometries from aggregated shapes and analyze each shape’s properties separately.

vector_patches = get_patches(vector_landscape, class_col = "class", direction = 4)
vector_patches
#> Simple feature collection with 40 features and 2 fields
#> Geometry type: POLYGON
#> Dimension:     XY
#> Bounding box:  xmin: 0 ymin: 0 xmax: 30 ymax: 30
#> CRS:           NA
#> First 10 features:
#>    class patch                       geometry
#> 1      1     1 POLYGON ((1 2, 0 2, 0 3, 0 ...
#> 2      1     2 POLYGON ((14 5, 15 5, 16 5,...
#> 3      1     3 POLYGON ((12 18, 12 19, 12 ...
#> 4      1     4 POLYGON ((10 19, 10 18, 9 1...
#> 5      1     5 POLYGON ((5 20, 6 20, 6 19,...
#> 6      1     6 POLYGON ((6 21, 7 21, 7 20,...
#> 7      1     7 POLYGON ((3 16, 4 16, 4 15,...
#> 8      1     8 POLYGON ((2 24, 2 23, 2 22,...
#> 9      1     9 POLYGON ((10 20, 11 20, 11 ...
#> 10     1    10 POLYGON ((13 24, 14 24, 14 ...

vector_patches |>
  dplyr::mutate(patch = as.factor(patch)) |>
  plot()

## Shape index
vm_p_shape(vector_patches, class_col = "class", patch_col = "patch")
#> # A tibble: 40 × 5
#>    level class id    metric value
#>    <chr> <chr> <chr> <chr>  <dbl>
#>  1 patch 1     1     shape   1.66
#>  2 patch 1     2     shape   1.37
#>  3 patch 1     3     shape   1.51
#>  4 patch 1     4     shape   1.41
#>  5 patch 1     5     shape   1.13
#>  6 patch 1     6     shape   1.13
#>  7 patch 1     7     shape   1.13
#>  8 patch 1     8     shape   1.15
#>  9 patch 1     9     shape   1.13
#> 10 patch 1     10    shape   1.13
#> # ℹ 30 more rows

## Number of patches
vm_c_np(vector_patches, class_col = "class")
#> # A tibble: 3 × 5
#>   level class id    metric value
#>   <chr> <chr> <chr> <chr>  <int>
#> 1 class 1     <NA>  np        19
#> 2 class 2     <NA>  np        14
#> 3 class 3     <NA>  np         7

## Mean squareness
vm_l_square_mn(vector_patches)
#> # A tibble: 1 × 5
#>   level     class id    metric value
#>   <chr>     <chr> <chr> <chr>  <dbl>
#> 1 landscape <NA>  <NA>  sq_mn  0.845

Another utility function is get_axes() which calculates the length of the major and minor axes of the shape. It is used to calculate the elongation metric in vm_p_elong() but since length of axes might be useful information itself, get_axes() was exported as a separate function.

get_axes(vector_patches, class_col = "class")
#> # A tibble: 40 × 6
#>    level class id    metric    major minor
#>    <chr> <chr> <chr> <chr>     <dbl> <dbl>
#>  1 patch 1     1     main_axes 10.8   5.28
#>  2 patch 1     2     main_axes  9.64  7.84
#>  3 patch 1     3     main_axes  8.38  5.16
#>  4 patch 1     4     main_axes  3.6   2.22
#>  5 patch 1     5     main_axes  1.42  1.42
#>  6 patch 1     6     main_axes  1.42  1.42
#>  7 patch 1     7     main_axes  1.42  1.42
#>  8 patch 1     8     main_axes  4.24  2.82
#>  9 patch 1     9     main_axes  1.42  1.42
#> 10 patch 1     10    main_axes  1.42  1.42
#> # ℹ 30 more rows

Contributing

This is an experimental version of the package, so any feedback and contributions in the form of pull requests are welcome.