Define default summary functions

Creates and returns a named list of default summary functions used throughout the analytics framework (e.g., by compute_fastsummary()). Each function is defined as a purrr-style formula (~) that operates on a vector .x and returns a scalar summary statistic. The returned list can be supplied directly to a summarization pipeline or extended by users with custom functions.

Usage

define_fns()

Format

The list includes the following summary functions:

sum: Sum of values, ignoring NAs.
mean: Arithmetic mean.
median: Median value.
cv: Coefficient of variation (requires a cv() helper).
cp_ratio: Custom "cp ratio" statistic (requires a cp_ratio() helper).
var: Sample variance.
iqr: Interquartile range, computed as diff(range(.x)).
min: Minimum value.
max: Maximum value.
count: Number of observations.
count_unique: Number of distinct (unique) values.
prop_na: Proportion of missing (NA) values.
prop_zero: Proportion of zero values among non-missing data.
p25: 25th percentile (first quartile).
p75: 75th percentile (third quartile).
p90: 90th percentile.
sd: Standard deviation.

Value

A named list of formula functions suitable for use with dplyr::across(), where each element name is the function label and the value is a one-sided formula that computes the summary.

Details

The returned list contains commonly used descriptive statistics for numeric vectors, including measures of central tendency, dispersion, distribution, and data quality (e.g., share of missing or zero values). Users can extend or override the defaults by appending their own named formulas before passing to compute_fastsummary().

Examples

# Load the default function set
fns <- define_fns()

# Inspect available summaries
names(fns)
#>  [1] "sum"          "mean"         "median"       "cv"           "cp_ratio"    
#>  [6] "var"          "iqr"          "min"          "max"          "count"       
#> [11] "prop"         "dtprop"       "count_unique" "prop_na"      "prop_zero"   
#> [16] "p25"          "p75"          "p90"          "sd"          

# Example usage with compute_fastsummary()
compute_fastsummary(
  data = tibble::tibble(
     country_code = c(rep("A", 100), rep("B", 100)),
     gross_salary_lcu = c(
      rnorm(100, mean = 1000, sd = 100),
      rnorm(100,  mean = 2000, sd = 100)
      ),
     net_salary_lcu = c(
      rnorm(100, mean = 0.7 * 1000, sd = 100),
      rnorm(100,  mean = 0.7 * 2000, sd = 100)
     )
  ) |> data.table::as.data.table(),
  cols = c("gross_salary_lcu", "net_salary_lcu"),
  groups = c("country_code"),
  fns = c("mean", "sd", "cv")
)
#> Warning: `groups` is deprecated and will be removed in a future release; use `group_cols` instead.
#>     country_code             indicator        value
#>           <char>                <fctr>        <num>
#>  1:            A gross_salary_lcu_mean 1.007079e+03
#>  2:            B gross_salary_lcu_mean 2.011486e+03
#>  3:            A   gross_salary_lcu_sd 1.050905e+02
#>  4:            B   gross_salary_lcu_sd 1.014279e+02
#>  5:            A   gross_salary_lcu_cv 1.043518e-01
#>  6:            B   gross_salary_lcu_cv 5.042437e-02
#>  7:            A   net_salary_lcu_mean 6.969953e+02
#>  8:            B   net_salary_lcu_mean 1.418860e+03
#>  9:            A     net_salary_lcu_sd 1.051104e+02
#> 10:            B     net_salary_lcu_sd 9.921035e+01
#> 11:            A     net_salary_lcu_cv 1.508051e-01
#> 12:            B     net_salary_lcu_cv 6.992258e-02

Usage

Format

Value

Details

See also

Examples