Prepare chart labels that should be repelled apart

This function is designed for plots, particularly dot plots, where some values are very close together and need to be pushed apart in order to avoid overlaps. It returns just the columns needed to know which points should be dodged apart.

It doesn't calculate new positions for those points— that depends on your specific use, but ggrepel::geom_text_repel is likely what you'll use with this. You'll probably call this to get a data frame of the points that should be dodged, then use dplyr::semi_join on the original data to plot the points that do need to be dodged, and dplyr::anti_join for the ones that don't need to be dodged (see example).

Usage

dodge_lbls(data, x, value, group, thresh, digits = 2, verbose = FALSE)

Arguments

data: A data frame
x: Bare name of the column being plotted on the independent axis
value: Bare name of the dependent variable
group: Bare column name of the grouping variable, likely used for point color
thresh: Numeric: the threshold value of the difference between points. Any pair of points with a difference of less than thresh will be considered too close, and will be included in the output as points to dodge.
digits: Number of digits to round to before calculating differences between points. Default: 2
verbose: Boolean: if TRUE, will report the number of observations being dodged. Defaults FALSE.

Value

A data frame of too-close points, with columns corresponding to the x and group columns.

Examples

if (requireNamespace("ggrepel", quietly = TRUE)) {
    library(ggplot2)

    cws24 <- cws_trend |>
        dplyr::filter(year == 2024)

    # some labels are overlapped & hard to read
    ggplot(cws24, aes(x = value, y = question)) +
        geom_point(aes(color = group), size = 6, alpha = 0.8) +
        geom_text(aes(label = percent100(value)), size = 3)

    to_dodge <- dodge_lbls(cws24,
        x = question, value = value, group = group,
        thresh = 0.01
    )

    # the first text geom will be observations that don't need to be dodged (anti-join)
    # the second will be ones that do need to be dodged (semi-join)
    ggplot(cws24, aes(x = value, y = question)) +
        geom_point(aes(color = group), size = 6, alpha = 0.8) +
        geom_text(aes(label = percent100(value)),
            size = 2.7,
            data = ~ dplyr::anti_join(., to_dodge, by = c("question", "group"))
        ) +
        ggrepel::geom_text_repel(aes(label = percent100(value), color = group),
            size = 2.7, seed = 1,
            data = ~ dplyr::semi_join(., to_dodge, by = c("question", "group")),
            direction = "x", nudge_y = 0.2
        )
}