Skip to contents

geom_plot and geom_plot_npc add ggplot objects as insets to the base ggplot, using syntax similar to that of geom_label and geom_text_s. In most respects they behave as any other ggplot geometry: they add a layer containing one or more grobs and grouping and faceting works as usual. The most common use of geom_plot is to add data labels that are themselves ggplots rather than text. geom_plot_npc is used to add ggplots as annotations to plots, but contrary to layer function annotate(), geom_plot_npc is data driven and respects grouping and facets, thus plot insets can differ among panels.


  mapping = NULL,
  data = NULL,
  stat = "identity",
  position = "identity",
  nudge_x = 0,
  nudge_y = 0,
  default.colour = "black",
  default.color = default.colour, = "box", =,
  default.alpha = 1, = "all",
  add.segments = TRUE,
  box.padding = 0.25,
  point.padding = 1e-06,
  segment.linewidth = 0.5,
  min.segment.length = 0,
  arrow = NULL,
  na.rm = FALSE,
  show.legend = FALSE,
  inherit.aes = FALSE

  mapping = NULL,
  data = NULL,
  stat = "identity",
  position = "identity",
  na.rm = FALSE,
  show.legend = FALSE,
  inherit.aes = FALSE



The aesthetic mapping, usually constructed with aes. Only needs to be set at the layer level if you are overriding the plot defaults.


A layer specific data set - only needed if you want to override the plot defaults.


The statistical transformation to use on the data for this layer, as a string.


Position adjustment, either as a string, or the result of a call to a position adjustment function.


other arguments passed on to layer. This can include aesthetics whose values you want to set, not map. See layer for more details.

nudge_x, nudge_y

Horizontal and vertical adjustments to nudge the starting position of each text label. The units for nudge_x and nudge_y are the same as for the data units on the x-axis and y-axis.

default.colour, default.color

A colour definition to use for elements not targeted by the colour aesthetic.,

A vector of character strings; "all", "text", "box" and "segment".


numeric in [0..1] A transparency value to use for elements not targeted by the alpha aesthetic.

A vector of character strings; "all", "text", "segment", "box", "box.line", and "box.fill".


logical Display connecting segments or arrows between original positions and displaced ones if both are available.

box.padding, point.padding

numeric By how much each end of the segments should shortened in mm.


numeric Width of the segments or arrows in mm.


numeric Segments shorter that the minimum length are not rendered, in mm.


specification for arrow heads, as created by arrow


If FALSE (the default), removes missing values with a warning. If TRUE silently removes missing values.


logical. Should this layer be included in the legends? NA, the default, includes if any aesthetics are mapped. FALSE never includes, and TRUE always includes.


If FALSE, overrides the default aesthetics, rather than combining with them. This is most useful for helper functions that define both data and aesthetics and shouldn't inherit behaviour from the default plot specification, e.g. borders.


A plot layer instance.


You can modify the size of inset plots with the vp.width and vp.height aesthetics. These can take a number between 0 (smallest possible inset) and 1 (whole plotting area width or height). The default value for for both of these aesthetics is 1/5. Thus, in contrast to geom_text and geom_text_s the size of the insets remains the same relative to the size of the plotting area irrespective of how the plot is rendered. The aspect ratio of insets is preserved and size is adjusted until the whole inset fits within the viewport.

By default this geom uses position_nudge_center which is backwards compatible with position_nudge but provides additional control on the direction of the nudging. In contrast to position_nudge, position_nudge_center and all other position functions defined in packages 'ggpp' and 'ggrepel' keep the original coordinates thus allowing the plotting of connecting segments and arrows.

This geom works only with tibbles as data, as its expects a list of ggplot objects ("gg" class) to be mapped to the label aesthetic.

The x and y aesthetics determine the position of the whole inset plot, similarly to that of a text label, justification is interpreted as indicating the position of the plot with respect to its x and y coordinates in the data, and angle is used to rotate the plot as a whole.

Of these two geoms only geom_plot supports the plotting of segments, as geom_plot_npc uses a coordinate system that is unrelated to data units and data.In the case of geom_plot_npc(), npcx and npcy aesthetics determine the position of the inset plot.


The insets are stored nested within the main ggplot object and contain their own copy of the data, and are rendered as grid grobs as normal ggplots at the time the main ggplot is rendered. They can have different themes.

Use annotate as redefined in 'ggpp' when adding insets as annotations (automatically available unless 'ggpp' is not attached). annotate cannot be used with the npcx and npcy pseudo-aesthetics.


You can modify text alignment with the vjust and hjust aesthetics. These can either be a number between 0 (right/bottom) and 1 (top/left) or a character ("left", "middle", "right", "bottom", "center", "top"). In addition, you can use special alignments for justification including "position", "inward" and "outward". Inward always aligns text towards the center of the plotting area, and outward aligns it away from the center of the plotting area. If tagged with _mean or _median (e.g., "outward_mean") the mean or median of the data in the panel along the corresponding axis is used as center. If the characters following the underscore represent a number (e.g., "outward_10.5") the reference point will be this value in data units. Position justification is computed based on the direction of the displacement of the position of the label so that each individual text or label is justified outwards from its original position. The default justification is "position".

If no position displacement is applied, or a position function defined in 'ggplot2' is used, these geometries behave similarly to the corresponding ones from package 'ggplot2' with a default justification of 0.5 and no segment drawn.

Position functions

Many layer functions from package 'ggpp' are designed to work seamlessly with position functions that keep, rather than discard, the original x and y positions in data when computing a new displaced position. See position_nudge_keep, position_dodge_keep, position_jitter_keep, position_nudge_center, position_nudge_line, position_nudge_to, position_dodgenudge, position_jitternudge, and position_stacknudge for examples and details of their use.

Plot boundaries and clipping

The "width" and "height" of an inset as for a text element are 0, so stacking and dodging inset plots will not work by default, and axis limits are not automatically expanded to include all inset plots. Obviously, insets do have height and width, but they are physical units, not data units. The amount of space they occupy on the main plot is not constant in data units of the base plot: when you modify scale limits, inset plots stay the same size relative to the physical size of the base plot.


The idea of implementing a geom_custom() for grobs has been discussed as an issue at

See also

Other geometries adding layers with insets: geom_table()


# inset plot with enlarged detail from a region of the main plot
p <-
  ggplot(data = mtcars, mapping = aes(wt, mpg)) +

df <- tibble(x = 0.01,
             y = 0.01,
             plot = list(p +
                         coord_cartesian(xlim = c(3, 4),
                                         ylim = c(13, 16)) +
                         labs(x = NULL, y = NULL) +
p +
  expand_limits(x = 0, y = 0) +
  geom_plot_npc(data = df,
                aes(npcx = x, npcy = y, label = plot))

p +
  expand_limits(x = 0, y = 0) +
  geom_plot_npc(data = df,
                aes(npcx = x, npcy = y, label = plot,
                vp.width = 1/2, vp.height = 1/4))

p +
  expand_limits(x = 0, y = 0) +
  geom_plot_npc(data = df,
                aes(npcx = x, npcy = y, label = plot),
                vp.width = 1/4, vp.height = 1/4)

p +
  geom_plot(data = df,
            aes(x = x + 3, y = y + 20, label = plot),
            nudge_x = -1, nudge_y = - 7,
            hjust = 0.5, vjust = 0.5,
            arrow = arrow(length = unit(0.5, "lines")),
            colour = "red",
            vp.width = 1/5, vp.height = 1/5)