geom-smooth.r

#' Smoothed conditional means
#'
#' Aids the eye in seeing patterns in the presence of overplotting.
#' `geom_smooth()` and `stat_smooth()` are effectively aliases: they
#' both use the same arguments. Use `stat_smooth()` if you want to
#' display the results with a non-standard geom.
#'
#' Calculation is performed by the (currently undocumented)
#' `predictdf()` generic and its methods.  For most methods the standard
#' error bounds are computed using the [predict()] method -- the
#' exceptions are `loess()`, which uses a t-based approximation, and
#' `glm()`, where the normal confidence interval is constructed on the link
#' scale and then back-transformed to the response scale.
#'
#' @eval rd_aesthetics("geom", "smooth")
#' @inheritParams layer
#' @inheritParams geom_point
#' @param geom,stat Use to override the default connection between
#'   `geom_smooth()` and `stat_smooth()`.
#' @seealso See individual modelling functions for more details:
#'   [lm()] for linear smooths,
#'   [glm()] for generalised linear smooths, and
#'   [loess()] for local smooths.
#' @export
#' @examples
#' ggplot(mpg, aes(displ, hwy)) +
#'   geom_point() +
#'   geom_smooth()
#'
#' # Use span to control the "wiggliness" of the default loess smoother.
#' # The span is the fraction of points used to fit each local regression:
#' # small numbers make a wigglier curve, larger numbers make a smoother curve.
#' ggplot(mpg, aes(displ, hwy)) +
#'   geom_point() +
#'   geom_smooth(span = 0.3)
#'
#' # Instead of a loess smooth, you can use any other modelling function:
#' ggplot(mpg, aes(displ, hwy)) +
#'   geom_point() +
#'   geom_smooth(method = lm, se = FALSE)
#'
#' ggplot(mpg, aes(displ, hwy)) +
#'   geom_point() +
#'   geom_smooth(method = lm, formula = y ~ splines::bs(x, 3), se = FALSE)
#'
#' # Smooths are automatically fit to each group (defined by categorical
#' # aesthetics or the group aesthetic) and for each facet.
#'
#' ggplot(mpg, aes(displ, hwy, colour = class)) +
#'   geom_point() +
#'   geom_smooth(se = FALSE, method = lm)
#' ggplot(mpg, aes(displ, hwy)) +
#'   geom_point() +
#'   geom_smooth(span = 0.8) +
#'   facet_wrap(~drv)
#'
#' \donttest{
#' binomial_smooth <- function(...) {
#'   geom_smooth(method = "glm", method.args = list(family = "binomial"), ...)
#' }
#' # To fit a logistic regression, you need to coerce the values to
#' # a numeric vector lying between 0 and 1.
#' ggplot(rpart::kyphosis, aes(Age, Kyphosis)) +
#'   geom_jitter(height = 0.05) +
#'   binomial_smooth()
#'
#' ggplot(rpart::kyphosis, aes(Age, as.numeric(Kyphosis) - 1)) +
#'   geom_jitter(height = 0.05) +
#'   binomial_smooth()
#'
#' ggplot(rpart::kyphosis, aes(Age, as.numeric(Kyphosis) - 1)) +
#'   geom_jitter(height = 0.05) +
#'   binomial_smooth(formula = y ~ splines::ns(x, 2))
#'
#' # But in this case, it's probably better to fit the model yourself
#' # so you can exercise more control and see whether or not it's a good model.
#' }
geom_smooth <- function(mapping = NULL, data = NULL,
                        stat = "smooth", position = "identity",
                        ...,
                        method = "auto",
                        formula = y ~ x,
                        se = TRUE,
                        na.rm = FALSE,
                        show.legend = NA,
                        inherit.aes = TRUE) {

  params <- list(
    na.rm = na.rm,
    se = se,
    ...
  )
  if (identical(stat, "smooth")) {
    params$method <- method
    params$formula <- formula
  }

  layer(
    data = data,
    mapping = mapping,
    stat = stat,
    geom = GeomSmooth,
    position = position,
    show.legend = show.legend,
    inherit.aes = inherit.aes,
    params = params
  )
}

#' @rdname ggplot2-ggproto
#' @format NULL
#' @usage NULL
#' @export
GeomSmooth <- ggproto("GeomSmooth", Geom,
  setup_data = function(data, params) {
    GeomLine$setup_data(data, params)
  },

  # The `se` argument is set to false here to make sure drawing the
  # geom and drawing the legend is in synch. If the geom is used by a
  # stat that doesn't set the `se` argument then `se` will be missing
  # and the legend key won't be drawn. With `se = FALSE` here the
  # ribbon won't be drawn either in that case, keeping the overall
  # behavior predictable and sensible. The user will realize that they
  # need to set `se = TRUE` to obtain the ribbon and the legend key.
  draw_group = function(data, panel_params, coord, se = FALSE) {
    ribbon <- transform(data, colour = NA)
    path <- transform(data, alpha = NA)

    has_ribbon <- se && !is.null(data$ymax) && !is.null(data$ymin)

    gList(
      if (has_ribbon) GeomRibbon$draw_group(ribbon, panel_params, coord),
      GeomLine$draw_panel(path, panel_params, coord)
    )
  },

  draw_key = draw_key_smooth,

  required_aes = c("x", "y"),
  optional_aes = c("ymin", "ymax"),

  default_aes = aes(colour = "#3366FF", fill = "grey60", size = 1,
    linetype = 1, weight = 1, alpha = 0.4)
)