Add reliability labels based on coefficient of variation

Adds a reliability column that classifies each row by its coefficient of variation (CV). There are two ways to call it:

Usage

add_reliability(data, cv_col = NULL, scale = c("percent", "proportion"))

Arguments

data: A data frame or tibble.
cv_col: The column of CV values to classify. If NULL (the default), the legacy _cv auto-detection path is used.
scale: Whether cv_col is on a "percent" (0-100, the default) or "proportion" (0-1) scale. Ignored on the legacy path.

Value

The input data frame with an additional reliability character column.

Details

Supply cv_col (recommended). The named column is classified with <= boundaries on the chosen scale. With scale = "percent" (the default) a CV of exactly 15 is "High" and exactly 30 is "Medium".
Omit cv_col (legacy). The single column ending in _cv is auto-detected and treated as a 0-1 proportion with strict < boundaries (High: CV < 0.15; Medium: 0.15 <= CV < 0.30; Low: CV >= 0.30).

Missing CV values yield NA. The result is a character column (not a factor).

Examples

# Recommended: name the CV column (percent scale)
df <- data.frame(location = c("A", "B", "C", "D"), cv = c(5, 15, 30, NA))
add_reliability(df, cv_col = cv)
#>   location cv reliability
#> 1        A  5        High
#> 2        B 15        High
#> 3        C 30      Medium
#> 4        D NA        <NA>

# Legacy: auto-detect a _cv column on the 0-1 proportion scale
df2 <- data.frame(location = c("A", "B", "C"), value_cv = c(0.05, 0.20, 0.35))
add_reliability(df2)
#>   location value_cv reliability
#> 1        A     0.05        High
#> 2        B     0.20      Medium
#> 3        C     0.35         Low