ecdfSteps

Compute Coordinates of an Empirical Distribution Function

ecdfSteps(x, extend)

Arguments

x

numeric vector, possibly with NAs that are ignored

extend

a 2-vector do extend the range of x (low, high). Set extend=FALSE to not extend x, or leave it missing to extend it 1/20th of the observed range on other side.

Value

a list with components x and y

Details

For a numeric vector uses the R built-in ecdf function to compute coordinates of the ECDF, with extension slightly below and above the range of x by default. This is useful for ggplot2 where the ECDF may need to be transformed. The returned object is suitable for creating stratified statistics using data.table and other methods.

See also

stats::ecdf()

Author

Frank Harrell

Examples

ecdfSteps(0:10)
#> $x
#>  [1] -0.5  0.0  1.0  2.0  3.0  4.0  5.0  6.0  7.0  8.0  9.0 10.0 10.5
#> 
#> $y
#>  [1] 0.00000000 0.09090909 0.18181818 0.27272727 0.36363636 0.45454545
#>  [7] 0.54545455 0.63636364 0.72727273 0.81818182 0.90909091 1.00000000
#> [13] 1.00000000
#> 
if (FALSE) { # \dontrun{
# Use data.table for obtaining ECDFs by country and region
w <- d[, ecdfSteps(z, extend=c(1,11)), by=.(country, region)]  # d is a DT
# Use ggplot2 to make one graph with multiple regions' ECDFs
# and use faceting for countries
ggplot(w, aes(x, y, color=region)) + geom_step() +
       facet_wrap(~ country)
} # }