round_by() rounds and leaves the values numeric, which is right when each
column holds one metric. format_by() formats each value to a string
instead, which is what a column of several metrics needs: a numeric
column spanning 1e-5 to 1e4 prints wholly in scientific notation, so a
count of 3 reads as 3.00e+00. DT::datatable() formats cell by cell
and so needs neither; knitr::kable() formats by column and so needs
format_by() for a comparison.
Arguments
- display
A list from
connectivity_display().
Examples
lizard <- habitat_connectivity(
habitat = example_habitat(),
barrier = example_barrier(),
species = "Blue Tongue Lizard",
interpatch_distance = 20,
verbose = FALSE
)
round_by(connectivity_display(lizard))
#> # A tibble: 1 × 8
#> Species `Distance (m)` Patches `Mesh (ha)` `P(connected)` `Mean area (m2)`
#> <chr> <dbl> <int> <dbl> <dbl> <dbl>
#> 1 Blue Tongu… 20 163 4.42 0.0000168 1612.
#> # ℹ 2 more variables: `Total area (ha)` <dbl>, `Resolution (m)` <chr>