Skip to content

ggplot2 3.5.0

This is a minor release that turned out quite beefy. It is focused on overhauling the guide system: the system responsible for displaying information from scales in the guise of axes and legends. As part of that overhaul, new guides have been implemented and existing guides have been refined. The look and feel of guides has been mostly preserved, but their internals and styling options have changed drastically.

Briefly summarising other highlights, we also welcome coord_radial() as a successor of coord_polar(). Initial support for newer graphical features, such as pattern fills has been added. The API has changed how I()/<AsIs> vectors interact with the scale system, namely: not at all.

Breaking changes

  • The guide system. As a whole. See ‘new features’ for more information. While the S3 guide generics are still in place, the S3 methods for guide_train(), guide_merge(), guide_geom(), guide_transform(), guide_gengrob() have been superseded by the respective ggproto methods. In practice, this will mean that NextMethod() or sub-classing ggplot2’s guides with the S3 system will no longer work.

  • By default, guide_legend() now only draws a key glyph for a layer when the value is in the layer’s data. To revert to the old behaviour, you can still set show.legend = c({aesthetic} = TRUE) (@teunbrand, #3648).

  • In the scale_{colour/fill}_gradient2() and scale_{colour/fill}_steps2() functions, the midpoint argument is transformed by the scale transformation (#3198).

  • The legend.key theme element is set to inherit from the panel.background theme element. The default themes no longer set the legend.key element. This causes a visual change with the default theme_gray() (#5549).

  • The scale_name argument in continuous_scale(), discrete_scale() and binned_scale() is soft-deprecated. If you have implemented custom scales, be advised to double-check that unnamed arguments ends up where they should (@teunbrand, #1312).

  • The legend.text.align and legend.title.align arguments in theme() are deprecated. The hjust setting of the legend.text and legend.title elements continues to fulfill the role of text alignment (@teunbrand, #5347).

  • ‘lines’ units in geom_label(), often used in the label.padding argument, are now are relative to the text size. This causes a visual change, but fixes a misalignment issue between the textbox and text (@teunbrand, #4753)

  • coord_flip() has been marked as superseded. The recommended alternative is to swap the x and y aesthetic and/or using the orientation argument in a layer (@teunbrand, #5130).

  • The trans argument in scales and secondary axes has been renamed to transform. The trans argument itself is deprecated. To access the transformation from the scale, a new get_transformation() method is added to Scale-classes (#5558).

  • Providing a numeric vector to theme(legend.position) has been deprecated. To set the default legend position inside the plot use theme(legend.position = "inside", legend.position.inside = c(...)) instead.

New features

  • Plot scales now ignore AsIs objects constructed with I(x), instead of invoking the identity scale. This allows these columns to co-exist with other layers that need a non-identity scale for the same aesthetic. Also, it makes it easy to specify relative positions (@teunbrand, #5142).

  • The fill aesthetic in many geoms now accepts grid’s patterns and gradients. For developers of layer extensions, this feature can be enabled by switching from fill = alpha(fill, alpha) to fill = fill_alpha(fill, alpha) when providing fills to grid::gpar() (@teunbrand, #3997).

  • New function check_device() for testing the availability of advanced graphics features introduced in R 4.1.0 onward (@teunbrand, #5332).

  • coord_radial() is a successor to coord_polar() with more customisation options. coord_radial() can:

    • integrate with the new guide system via a dedicated guide_axis_theta() to display the angle coordinate.
    • in addition to drawing full circles, also draw circle sectors by using the end argument.
    • avoid data vanishing in the center of the plot by setting the donut argument.
    • adjust the angle aesthetic of layers, such as geom_text(), to align with the coordinate system using the rotate_angle argument.

The guide system

The guide system encompassing axes and legends, as the last remaining chunk of ggplot2, has been rewritten to use the <ggproto> system instead of the S3 system. This change was a necessary step to officially break open the guide system for extension package developers. The axes and legends now inherit from a <Guide> class, which makes them extensible in the same manner as geoms, stats, facets and coords (#3329, @teunbrand)

  • The most user-facing change is that the styling of guides is rewired through the theme system. Guides now have a theme argument that can style individual guides, while theme() has gained additional arguments to style guides. Theme elements declared in the guide override theme elements set through the plot. The new theme elements for guides are: legend.key.spacing{.x/.y}, legend.frame, legend.axis.line, legend.ticks, legend.ticks.length, legend.text.position and legend.title.position. Previous style options in the arguments of guide_*() functions are soft-deprecated.

  • Unfortunately, we could not fully preserve the function of pre-existing guide extensions written in the S3 system. A fallback for these old guides is encapsulated in the <GuideOld> class, which calls the old S3 generics. The S3 methods have been removed as part of cleaning up, so the old guides will still work if the S3 methods are reimplemented, but we encourage to switch to the new system (#2728).

  • The order argument of guides now strictly needs to be a length-1 integer (#4958).

Axes
  • New guide_axis_stack() to combine other axis guides on top of one another.

  • New guide_axis_theta() to draw an axis in a circular arc in coord_radial(). The guide can be controlled by adding guides(theta = guide_axis_theta(...)) to a plot.

  • New guide_axis_logticks() can be used to draw logarithmic tick marks as an axis. It supersedes the annotation_logticks() function (@teunbrand, #5325).

  • guide_axis() gains a minor.ticks argument to draw minor ticks (#4387).

  • guide_axis() gains a cap argument that can be used to trim the axis line to extreme breaks (#4907).

  • Primary axis titles are now placed at the primary guide, so that guides(x = guide_axis(position = "top")) will display the title at the top by default (#4650).

  • The default vjust for the axis.title.y.right element is now 1 instead of

  • Unknown secondary axis guide positions are now inferred as the opposite of the primary axis guide when the latter has a known position (#4650).

Legends
  • New guide_custom() function for drawing custom graphical objects (grobs) unrelated to scales in legend positions (#5416).

  • All legends have acquired a position argument, that allows individual guides to deviate from the legend.position set in the theme() function. This means that legends can now be placed at multiple sides of the plot (#5488).

  • The spacing between legend keys and their labels, in addition to legends and their titles, is now controlled by the text’s margin setting. Not specifying margins will automatically add appropriate text margins. To control the spacing within a legend between keys, the new legend.key.spacing.{x/y} argument can be used in theme(). This leaves the legend.spacing theme setting dedicated to solely controlling the spacing between different guides (#5455).

  • guide_colourbar() and guide_coloursteps() gain an alpha argument to set the transparency of the bar (#5085).

  • New display argument in guide_colourbar() supplants the raster argument. In R 4.1.0 and above, display = "gradient" will draw a gradient.

  • Legend keys that can draw arrows have their size adjusted for arrows.

  • When legend titles are larger than the legend, title justification extends to the placement of keys and labels (#1903).

  • Glyph drawing functions of the draw_key_*() family can now set "width" and "height" attributes (in centimetres) to the produced keys to control their displayed size in the legend.

  • coord_sf() now uses customisable guides provided in the scales or guides() function (@teunbrand).

Improvements

  • guide_coloursteps(even.steps = FALSE) now draws one rectangle per interval instead of many small ones (#5481).

  • draw_key_label() now better reflects the appearance of labels (#5561).

  • position_stack() no longer silently removes missing data, which is now handled by the geom instead of position (#3532).

  • The minor_breaks function argument in scales can now also take a function with two arguments: the scale’s limits and the scale’s major breaks (#3583).

  • Failing to fit or predict in stat_smooth() now gives a warning and omits the failed group, instead of throwing an error (@teunbrand, #5352).

  • labeller() now handles unspecified entries from lookup tables (@92amartins, #4599).

  • fortify.default() now accepts a data-frame-like object granted the object exhibits healthy dim(), colnames(), and as.data.frame() behaviours (@hpages, #5390).

  • geom_violin() gains a bounds argument analogous to geom_density()s (@eliocamp, #5493).

  • To apply dodging more consistently in violin plots, stat_ydensity() now has a drop argument to keep or discard groups with 1 observation.

  • geom_boxplot() gains a new argument, staplewidth that can draw staples at the ends of whiskers (@teunbrand, #5126)

  • geom_boxplot() gains an outliers argument to switch outliers on or off, in a manner that does affects the scale range. For hiding outliers that does not affect the scale range, you can continue to use outlier.shape = NA (@teunbrand, #4892).

  • Nicer error messages for xlim/ylim arguments in coord-* functions (@92amartins, #4601, #5297).

  • You can now omit either xend or yend from geom_segment() as only one of these is now required. If one is missing, it will be filled from the x and y aesthetics respectively. This makes drawing horizontal or vertical segments a little bit more convenient (@teunbrand, #5140).

  • When geom_path() has aesthetics varying within groups, the arrow() is applied to groups instead of individual segments (@teunbrand, #4935).

  • geom_text() and geom_label() gained a size.unit parameter that set the text size to millimetres, points, centimetres, inches or picas (@teunbrand, #3799).

  • geom_label() now uses the angle aesthetic (@teunbrand, #2785)

  • The label.padding argument in geom_label() now supports inputs created with the margin() function (#5030).

  • ScaleContinuous$get_breaks() now only calls scales::zero_range() on limits in transformed space, rather than in data space (#5304).

  • Scales throw more informative messages (@teunbrand, #4185, #4258)

  • scale_*_manual() with a named values argument now emits a warning when none of those names match the values found in the data (@teunbrand, #5298).

  • The name argument in most scales is now explicitly the first argument (#5535)

  • The translate_shape_string() internal function is now exported for use in extensions of point layers (@teunbrand, #5191).

  • To improve width calculation in bar plots with empty factor levels, resolution() considers mapped_discrete values as having resolution 1 (@teunbrand, #5211)

  • In theme(), some elements can be specified with rel() to inherit from unit-class objects in a relative fashion (@teunbrand, #3951).

  • theme() now supports splicing a list of arguments (#5542).

  • In the theme element hierarchy, parent elements that are a strict subclass of child elements now confer their subclass upon the children (#5457).

  • New plot.tag.location in theme() can control placement of the plot tag in the "margin", "plot" or the new "panel" option (#4297).

  • coord_munch() can now close polygon shapes (@teunbrand, #3271)

  • Aesthetics listed in geom_*() and stat_*() layers now point to relevant documentation (@teunbrand, #5123).

  • The new argument axes in facet_grid() and facet_wrap() controls the display of axes at interior panel positions. Additionally, the axis.labels argument can be used to only draw tick marks or fully labelled axes (@teunbrand, #4064).

  • coord_polar() can have free scales in facets (@teunbrand, #2815).

  • The get_guide_data() function can be used to extract position and label information from the plot (#5004).

  • Improve performance of layers without positional scales (@zeehio, #4990)

  • More informative error for mismatched direction/theme(legend.direction = ...) arguments (#4364, #4930).

Bug fixes

Internal changes

  • The ScaleContinuous$get_breaks() method no longer censors the computed breaks.

  • The ggplot object now contains $layout which points to the Layout ggproto object and will be used by the ggplot_build.ggplot method. This was exposed so that package developers may extend the behaviour of the Layout ggproto object without needing to develop an entirely new ggplot_build method (@jtlandis, #5077).

  • Guide building is now part of ggplot_build() instead of ggplot_gtable() to allow guides to observe unmapped data (#5483).

  • The titleGrob() function has been refactored to be faster and less complicated.

  • The scales_*() functions related to managing the <ScalesList> class have been implemented as methods in the <ScalesList> class, rather than stray functions (#1310).

ggplot2 3.4.4

CRAN release: 2023-10-12

This hotfix release adapts to a change in r-devel’s base::is.atomic() and the upcoming retirement of maptools.

ggplot2 3.4.3

CRAN release: 2023-08-14

This hotfix release addresses a version comparison change in r-devel. There are no user-facing or breaking changes.

ggplot2 3.4.2

CRAN release: 2023-04-03

This is a hotfix release anticipating changes in r-devel, but folds in upkeep changes and a few bug fixes as well.

Minor improvements

Bug fixes

ggplot2 3.4.1

CRAN release: 2023-02-10

This is a small release focusing on fixing regressions in the 3.4.0 release and minor polishes.

Breaking changes

New features

Bug fixes

In binned scales:

  • Automatic breaks should no longer be out-of-bounds, and automatic limits are adjusted to include breaks (@teunbrand, #5082).

  • Zero-range limits no longer throw an error and are treated akin to continuous scales with zero-range limits (@teunbrand, #5066).

  • The trans = "date" and trans = "time" transformations were made compatible (@teunbrand, #4217).

ggplot2 3.4.0

CRAN release: 2022-11-04

This is a minor release focusing on tightening up the internals and ironing out some inconsistencies in the API. The biggest change is the addition of the linewidth aesthetic that takes of sizing the width of any line from size. This change, while attempting to be as non-breaking as possible, has the potential to change the look of some of your plots.

Other notable changes is a complete redo of the error and warning messaging in ggplot2 using the cli package. Messaging is now better contextualised and it should be easier to identify which layer an error is coming from. Last, we have now made the switch to using the vctrs package internally which means that support for vctrs classes as variables should improve, along with some small gains in rendering speed.

Breaking changes

  • A linewidth aesthetic has been introduced and supersedes the size aesthetic for scaling the width of lines in line based geoms. size will remain functioning but deprecated for these geoms and it is recommended to update all code to reflect the new aesthetic. For geoms that have both point sizing and linewidth sizing (geom_pointrange() and geom_sf) size now only refers to sizing of points which can leads to a visual change in old code (@thomasp85, #3672)

  • The default line width for polygons in geom_sf() have been decreased to 0.2 to reflect that this is usually used for demarking borders where a thinner line is better suited. This change was made since we already induced a visual change in geom_sf() with the introduction of the linewidth aesthetic.

  • The dot-dot notation (..var..) and stat(), which have been superseded by after_stat(), are now formally deprecated (@yutannihilation, #3693).

  • qplot() is now formally deprecated (@yutannihilation, #3956).

  • stage() now properly refers to the values without scale transformations for the stage of after_stat. If your code requires the scaled version of the values for some reason, you have to apply the same transformation by yourself, e.g. sqrt() for scale_{x,y}_sqrt() (@yutannihilation and @teunbrand, #4155).

  • Use rlang::hash() instead of digest::digest(). This update may lead to changes in the automatic sorting of legends. In order to enforce a specific legend order use the order argument in the guide. (@thomasp85, #4458)

  • referring to x in backquoted expressions with label_bquote() is no longer possible.

  • The ticks.linewidth and frame.linewidth parameters of guide_colourbar() are now multiplied with .pt like elsewhere in ggplot2. It can cause visual changes when these arguments are not the defaults and these changes can be restored to their previous behaviour by adding / .pt (@teunbrand #4314).

  • scale_*_viridis_b() now uses the full range of the viridis scales (@gregleleu, #4737)

New features

  • geom_col() and geom_bar() gain a new just argument. This is set to 0.5 by default; use just = 0/just = 1 to place columns on the left/right of the axis breaks. (@wurli, #4899)

  • geom_density() and stat_density() now support bounds argument to estimate density with boundary correction (@echasnovski, #4013).

  • ggplot now checks during statistical transformations whether any data columns were dropped and warns about this. If stats intend to drop data columns they can declare them in the new field dropped_aes. (@clauswilke, #3250)

  • ... supports rlang::list2 dynamic dots in all public functions. (@mone27, #4764)

  • theme() now has a strip.clip argument, that can be set to "off" to prevent the clipping of strip text and background borders (@teunbrand, #4118)

  • geom_contour() now accepts a function in the breaks argument (@eliocamp, #4652).

Minor improvements and bug fixes

ggplot2 3.3.6

CRAN release: 2022-05-03

This is a very small release only applying an internal change to comply with R 4.2 and its deprecation of default.stringsAsFactors(). There are no user facing changes and no breaking changes.

ggplot2 3.3.5

CRAN release: 2021-06-25

This is a very small release focusing on fixing a couple of untenable issues that surfaced with the 3.3.4 release

ggplot2 3.3.4

CRAN release: 2021-06-16

This is a larger patch release fixing a huge number of bugs and introduces a small selection of feature refinements.

Features

  • Alt-text can now be added to a plot using the alt label, i.e + labs(alt = ...). Currently this alt text is not automatically propagated, but we plan to integrate into Shiny, RMarkdown, and other tools in the future. (@thomasp85, #4477)

  • Add support for the BrailleR package for creating descriptions of the plot when rendered (@thomasp85, #4459)

  • coord_sf() now has an argument default_crs that specifies the coordinate reference system (CRS) for non-sf layers and scale/coord limits. This argument defaults to NULL, which means non-sf layers are assumed to be in projected coordinates, as in prior ggplot2 versions. Setting default_crs = sf::st_crs(4326) provides a simple way to interpret x and y positions as longitude and latitude, regardless of the CRS used by coord_sf(). Authors of extension packages implementing stat_sf()-like functionality are encouraged to look at the source code of stat_sf()’s compute_group() function to see how to provide scale-limit hints to coord_sf() (@clauswilke, #3659).

  • ggsave() now uses ragg to render raster output if ragg is available. It also handles custom devices that sets a default unit (e.g. ragg::agg_png) correctly (@thomasp85, #4388)

  • ggsave() now returns the saved file location invisibly (#3379, @eliocamp). Note that, as a side effect, an unofficial hack <ggplot object> + ggsave() no longer works (#4513).

  • The scale arguments limits, breaks, minor_breaks, labels, rescaler and oob now accept purrr style lambda notation (@teunbrand, #4427). The same is true for as_labeller() (and therefore also labeller()) (@netique, #4188).

  • Manual scales now allow named vectors passed to values to contain fewer elements than existing in the data. Elements not present in values will be set to NA (@thomasp85, #3451)

  • Date and datetime position scales support out-of-bounds (oob) arguments to control how limits affect data outside those limits (@teunbrand, #4199).

Fixes

ggplot2 3.3.3

CRAN release: 2020-12-30

This is a small patch release mainly intended to address changes in R and CRAN. It further changes the licensing model of ggplot2 to an MIT license.

ggplot2 3.3.2

CRAN release: 2020-06-19

This is a small release focusing on fixing regressions introduced in 3.3.1.

ggplot2 3.3.1

CRAN release: 2020-05-28

This is a small release with no code change. It removes all malicious links to a site that got hijacked from the readme and pkgdown site.

ggplot2 3.3.0

CRAN release: 2020-03-05

This is a minor release but does contain a range of substantial new features, along with the standard bug fixes. The release contains a few visual breaking changes, along with breaking changes for extension developers due to a shift in internal representation of the position scales and their axes. No user breaking changes are included.

This release also adds Dewey Dunnington (@paleolimbot) to the core team.

Breaking changes

There are no user-facing breaking changes, but a change in some internal representations that extension developers may have relied on, along with a few breaking visual changes which may cause visual tests in downstream packages to fail.

  • The panel_params field in the Layout now contains a list of list of ViewScale objects, describing the trained coordinate system scales, instead of the list object used before. Any extensions that use this field will likely break, as will unit tests that checks aspects of this.

  • element_text() now issues a warning when vectorized arguments are provided, as in colour = c("red", "green", "blue"). Such use is discouraged and not officially supported (@clauswilke, #3492).

  • Changed theme_grey() setting for legend key so that it creates no border (NA) rather than drawing a white one. (@annennenne, #3180)

  • geom_ribbon() now draws separate lines for the upper and lower intervals if colour is mapped. Similarly, geom_area() and geom_density() now draw the upper lines only in the same case by default. If you want old-style full stroking, use outline.type = "full" (@yutannihilation, #3503 / @thomasp85, #3708).

New features

  • The evaluation time of aesthetics can now be controlled to a finer degree. after_stat() supersedes the use of stat() and ..var..-notation, and is joined by after_scale() to allow for mapping to scaled aesthetic values. Remapping of the same aesthetic is now supported with stage(), so you can map a data variable to a stat aesthetic, and remap the same aesthetic to something else after statistical transformation (@thomasp85, #3534)

  • All coord_*() functions with xlim and ylim arguments now accept vectors with NA as a placeholder for the minimum or maximum value (e.g., ylim = c(0, NA) would zoom the y-axis from 0 to the maximum value observed in the data). This mimics the behaviour of the limits argument in continuous scale functions (@paleolimbot, #2907).

  • Allowed reversing of discrete scales by re-writing get_limits() (@AnneLyng, #3115)

  • All geoms and stats that had a direction (i.e. where the x and y axes had different interpretation), can now freely choose their direction, instead of relying on coord_flip(). The direction is deduced from the aesthetic mapping, but can also be specified directly with the new orientation argument (@thomasp85, #3506).

  • Position guides can now be customized using the new guide_axis(), which can be passed to position scale_*() functions or via guides(). The new axis guide (guide_axis()) comes with arguments check.overlap (automatic removal of overlapping labels), angle (easy rotation of axis labels), and n.dodge (dodge labels into multiple rows/columns) (@paleolimbot, #3322).

  • A new scale type has been added, that allows binning of aesthetics at the scale level. It has versions for both position and non-position aesthetics and comes with two new guides (guide_bins and guide_coloursteps) (@thomasp85, #3096)

  • scale_x_continuous() and scale_y_continuous() gains an n.breaks argument guiding the number of automatic generated breaks (@thomasp85, #3102)

  • Added stat_contour_filled() and geom_contour_filled(), which compute and draw filled contours of gridded data (@paleolimbot, #3044). geom_contour() and stat_contour() now use the isoband package to compute contour lines. The complete parameter (which was undocumented and has been unused for at least four years) was removed (@paleolimbot, #3044).

  • Themes have gained two new parameters, plot.title.position and plot.caption.position, that can be used to customize how plot title/subtitle and plot caption are positioned relative to the overall plot (@clauswilke, #3252).

Extensions

  • Geom now gains a setup_params() method in line with the other ggproto classes (@thomasp85, #3509)

  • The newly added function register_theme_elements() now allows developers of extension packages to define their own new theme elements and place them into the ggplot2 element tree (@clauswilke, #2540).

Minor improvements and bug fixes

ggplot2 3.2.1

CRAN release: 2019-08-10

This is a patch release fixing a few regressions introduced in 3.2.0 as well as fixing some unit tests that broke due to upstream changes.

  • position_stack() no longer changes the order of the input data. Changes to the internal behaviour of geom_ribbon() made this reordering problematic with ribbons that spanned y = 0 (#3471)
  • Using qplot() with a single positional aesthetic will no longer title the non-specified scale as "NULL" (#3473)
  • Fixes unit tests for sf graticule labels caused by changes to sf

ggplot2 3.2.0

CRAN release: 2019-06-16

This is a minor release with an emphasis on internal changes to make ggplot2 faster and more consistent. The few interface changes will only affect the aesthetics of the plot in minor ways, and will only potentially break code of extension developers if they have relied on internals that have been changed. This release also sees the addition of Hiroaki Yutani (@yutannihilation) to the core developer team.

With the release of R 3.6, ggplot2 now requires the R version to be at least 3.2, as the tidyverse is committed to support 5 major versions of R.

Breaking changes

  • Two patches (#2996 and #3050) fixed minor rendering problems. In most cases, the visual changes are so subtle that they are difficult to see with the naked eye. However, these changes are detected by the vdiffr package, and therefore any package developers who use vdiffr to test for visual correctness of ggplot2 plots will have to regenerate all reference images.

  • In some cases, ggplot2 now produces a warning or an error for code that previously produced plot output. In all these cases, the previous plot output was accidental, and the plotting code uses the ggplot2 API in a way that would lead to undefined behavior. Examples include a missing group aesthetic in geom_boxplot() (#3316), annotations across multiple facets (#3305), and not using aesthetic mappings when drawing ribbons with geom_ribbon() (#3318).

New features

  • This release includes a range of internal changes that speeds up plot generation. None of the changes are user facing and will not break any code, but in general ggplot2 should feel much faster. The changes includes, but are not limited to:

    • Caching ascent and descent dimensions of text to avoid recalculating it for every title.

    • Using a faster data.frame constructor as well as faster indexing into data.frames

    • Removing the plyr dependency, replacing plyr functions with faster equivalents.

  • geom_polygon() can now draw polygons with holes using the new subgroup aesthetic. This functionality requires R 3.6.0 (@thomasp85, #3128)

  • Aesthetic mappings now accept functions that return NULL (@yutannihilation, #2997).

  • stat_function() now accepts rlang/purrr style anonymous functions for the fun parameter (@dkahle, #3159).

  • geom_rug() gains an “outside” option to allow for moving the rug tassels to outside the plot area (@njtierney, #3085) and a length option to allow for changing the length of the rug lines (@daniel-wells, #3109).

  • All geoms now take a key_glyph paramter that allows users to customize how legend keys are drawn (@clauswilke, #3145). In addition, a new key glyph timeseries is provided to draw nice legends for time series (@mitchelloharawild, #3145).

Extensions

  • Layers now have a new member function setup_layer() which is called at the very beginning of the plot building process and which has access to the original input data and the plot object being built. This function allows the creation of custom layers that autogenerate aesthetic mappings based on the input data or that filter the input data in some form. For the time being, this feature is not exported, but it has enabled the development of a new layer type, layer_sf() (see next item). Other special-purpose layer types may be added in the future (@clauswilke, #2872).

  • A new layer type layer_sf() can auto-detect and auto-map sf geometry columns in the data. It should be used by extension developers who are writing new sf-based geoms or stats (@clauswilke, #3232).

  • x0 and y0 are now recognized positional aesthetics so they will get scaled if used in extension geoms and stats (@thomasp85, #3168)

  • Continuous scale limits now accept functions which accept the default limits and return adjusted limits. This makes it possible to write a function that e.g. ensures the limits are always a multiple of 100, regardless of the data (@econandrew, #2307).

Minor improvements and bug fixes

ggplot2 3.1.0

CRAN release: 2018-10-25

Breaking changes

This is a minor release and breaking changes have been kept to a minimum. End users of ggplot2 are unlikely to encounter any issues. However, there are a few items that developers of ggplot2 extensions should be aware of. For additional details, see also the discussion accompanying issue #2890.

  • In non-user-facing internal code (specifically in the aes() function and in the aesthetics argument of scale functions), ggplot2 now always uses the British spelling for aesthetics containing the word “colour”. When users specify a “color” aesthetic it is automatically renamed to “colour”. This renaming is also applied to non-standard aesthetics that contain the word “color”. For example, “point_color” is renamed to “point_colour”. This convention makes it easier to support both British and American spelling for novel, non-standard aesthetics, but it may require some adjustment for packages that have previously introduced non-standard color aesthetics using American spelling. A new function standardise_aes_names() is provided in case extension writers need to perform this renaming in their own code (@clauswilke, #2649).

  • Functions that generate other functions (closures) now force the arguments that are used from the generated functions, to avoid hard-to-catch errors. This may affect some users of manual scales (such as scale_colour_manual(), scale_fill_manual(), etc.) who depend on incorrect behavior (@krlmlr, #2807).

  • Coord objects now have a function backtransform_range() that returns the panel range in data coordinates. This change may affect developers of custom coords, who now should implement this function. It may also affect developers of custom geoms that use the range() function. In some applications, backtransform_range() may be more appropriate (@clauswilke, #2821).

New features

  • coord_sf() has much improved customization of axis tick labels. Labels can now be set manually, and there are two new parameters, label_graticule and label_axes, that can be used to specify which graticules to label on which side of the plot (@clauswilke, #2846, #2857, #2881).

  • Two new geoms geom_sf_label() and geom_sf_text() can draw labels and text on sf objects. Under the hood, a new stat_sf_coordinates() calculates the x and y coordinates from the coordinates of the sf geometries. You can customize the calculation method via fun.geometry argument (@yutannihilation, #2761).

Minor improvements and fixes

ggplot2 3.0.0

CRAN release: 2018-07-03

Breaking changes

  • ggplot2 now supports/uses tidy evaluation (as described below). This is a major change and breaks a number of packages; we made this breaking change because it is important to make ggplot2 more programmable, and to be more consistent with the rest of the tidyverse. The best general (and detailed) introduction to tidy evaluation can be found in the meta programming chapters in Advanced R.

    The primary developer facing change is that aes() now contains quosures (expression + environment pairs) rather than symbols, and you’ll need to take a different approach to extracting the information you need. A common symptom of this change are errors “undefined columns selected” or “invalid ‘type’ (list) of argument” (#2610). As in the previous version, constants (like aes(x = 1) or aes(colour = "smoothed")) are stored as is.

    In this version of ggplot2, if you need to describe a mapping in a string, use quo_name() (to generate single-line strings; longer expressions may be abbreviated) or quo_text() (to generate non-abbreviated strings that may span multiple lines). If you do need to extract the value of a variable instead use rlang::eval_tidy(). You may want to condition on (packageVersion("ggplot2") <= "2.2.1") so that your code can work with both released and development versions of ggplot2.

    We recognise that this is a big change and if you’re not already familiar with rlang, there’s a lot to learn. If you are stuck, or need any help, please reach out on https://community.rstudio.com.

  • Error: Column y must be a 1d atomic vector or a list

    Internally, ggplot2 now uses as.data.frame(tibble::as_tibble(x)) to convert a list into a data frame. This improves ggplot2’s support for list-columns (needed for sf support), at a small cost: you can no longer use matrix-columns. Note that unlike tibble we still allow column vectors such as returned by base::scale() because of their widespread use.

  • Error: More than one expression parsed

    Previously aes_string(x = c("a", "b", "c")) silently returned aes(x = a). Now this is a clear error.

  • Error: data must be uniquely named but has duplicate columns

    If layer data contains columns with identical names an error will be thrown. In earlier versions the first occurring column was chosen silently, potentially masking that the wrong data was chosen.

  • Error: Aesthetics must be either length 1 or the same as the data

    Layers are stricter about the columns they will combine into a single data frame. Each aesthetic now must be either the same length as the data frame or a single value. This makes silent recycling errors much less likely.

  • Error: coord_* doesn’t support free scales

    Free scales only work with selected coordinate systems; previously you’d get an incorrect plot.

  • Error in f(…) : unused argument (range = c(0, 1))

    This is because the oob argument to scale has been set to a function that only takes a single argument; it needs to take two arguments (x, and range).

  • Error: unused argument (output)

    The function guide_train() now has an optional parameter aesthetic that allows you to override the aesthetic setting in the scale. To make your code work with the both released and development versions of ggplot2 appropriate, add aesthetic = NULL to the guide_train() method signature.

    
    # old
    guide_train.legend <- function(guide, scale) {...}
    
    # new 
    guide_train.legend <- function(guide, scale, aesthetic = NULL) {...}

    Then, inside the function, replace scale$aesthetics[1], aesthetic %||% scale$aesthetics[1]. (The %||% operator is defined in the rlang package).

    
    # old
    setNames(list(scale$map(breaks)), scale$aesthetics[1])
    
    # new
    setNames(list(scale$map(breaks)), aesthetic %||% scale$aesthetics[1])
  • The long-deprecated subset argument to layer() has been removed.

Tidy evaluation

  • aes() now supports quasiquotation so that you can use !!, !!!, and :=. This replaces aes_() and aes_string() which are now soft-deprecated (but will remain around for a long time).

  • facet_wrap() and facet_grid() now support vars() inputs. Like dplyr::vars(), this helper quotes its inputs and supports quasiquotation. For instance, you can now supply faceting variables like this: facet_wrap(vars(am, cyl)) instead of facet_wrap(~am + cyl). Note that the formula interface is not going away and will not be deprecated. vars() is simply meant to make it easier to create functions around facet_wrap() and facet_grid().

    The first two arguments of facet_grid() become rows and cols and now support vars() inputs. Note however that we took special care to ensure complete backward compatibility. With this change facet_grid(vars(cyl), vars(am, vs)) is equivalent to facet_grid(cyl ~ am + vs), and facet_grid(cols = vars(am, vs)) is equivalent to facet_grid(. ~ am + vs).

    One nice aspect of the new interface is that you can now easily supply names: facet_grid(vars(Cylinder = cyl), labeller = label_both) will give nice label titles to the facets. Of course, those names can be unquoted with the usual tidy eval syntax.

sf

  • ggplot2 now has full support for sf with geom_sf() and coord_sf():

    
    nc <- sf::st_read(system.file("shape/nc.shp", package = "sf"), quiet = TRUE)
    ggplot(nc) +
      geom_sf(aes(fill = AREA))

    It supports all simple features, automatically aligns CRS across layers, sets up the correct aspect ratio, and draws a graticule.

New features

  • ggplot2 now works on R 3.1 onwards, and uses the vdiffr package for visual testing.

  • In most cases, accidentally using %>% instead of + will generate an informative error (#2400).

  • New syntax for calculated aesthetics. Instead of using aes(y = ..count..) you can (and should!) use aes(y = stat(count)). stat() is a real function with documentation which hopefully will make this part of ggplot2 less confusing (#2059).

    stat() is particularly nice for more complex calculations because you only need to specify it once: aes(y = stat(count / max(count))), rather than aes(y = ..count.. / max(..count..))

  • New tag label for adding identification tags to plots, typically used for labelling a subplot with a letter. Add a tag with labs(tag = "A"), style it with the plot.tag theme element, and control position with the plot.tag.position theme setting (@thomasp85).

Layers: geoms, stats, and position adjustments

  • geom_segment() and geom_curve() have a new arrow.fill parameter which allows you to specify a separate fill colour for closed arrowheads (@hrbrmstr and @clauswilke, #2375).

  • geom_point() and friends can now take shapes as strings instead of integers, e.g. geom_point(shape = "diamond") (@daniel-barnett, #2075).

  • position_dodge() gains a preserve argument that allows you to control whether the total width at each x value is preserved (the current default), or ensure that the width of a single element is preserved (what many people want) (#1935).

  • New position_dodge2() provides enhanced dodging for boxplots. Compared to position_dodge(), position_dodge2() compares xmin and xmax values
    to determine which elements overlap, and spreads overlapping elements evenly within the region of overlap. position_dodge2() is now the default position adjustment for geom_boxplot(), because it handles varwidth = TRUE, and will be considered for other geoms in the future.

    The padding parameter adds a small amount of padding between elements (@karawoo, #2143) and a reverse parameter allows you to reverse the order of placement (@karawoo, #2171).

  • New stat_qq_line() makes it easy to add a simple line to a Q-Q plot, which makes it easier to judge the fit of the theoretical distribution (@nicksolomon).

Scales and guides

  • Improved support for mapping date/time variables to alpha, size, colour, and fill aesthetics, including date_breaks and date_labels arguments (@karawoo, #1526), and new scale_alpha() variants (@karawoo, #1526).

  • Improved support for ordered factors. Ordered factors throw a warning when mapped to shape (unordered factors do not), and do not throw warnings when mapped to size or alpha (unordered factors do). Viridis is used as the default colour and fill scale for ordered factors (@karawoo, #1526).

  • The expand argument of scale_*_continuous() and scale_*_discrete() now accepts separate expansion values for the lower and upper range limits. The expansion limits can be specified using the convenience function expand_scale().

    Separate expansion limits may be useful for bar charts, e.g. if one wants the bottom of the bars to be flush with the x axis but still leave some (automatically calculated amount of) space above them:

    
    ggplot(mtcars) +
        geom_bar(aes(x = factor(cyl))) +
        scale_y_continuous(expand = expand_scale(mult = c(0, .1)))

    It can also be useful for line charts, e.g. for counts over time, where one wants to have a ’hard’ lower limit of y = 0 but leave the upper limit unspecified (and perhaps differing between panels), with some extra space above the highest point on the line (with symmetrical limits, the extra space above the highest point could in some cases cause the lower limit to be negative).

    The old syntax for the expand argument will, of course, continue to work (@huftis, #1669).

  • scale_colour_continuous() and scale_colour_gradient() are now controlled by global options ggplot2.continuous.colour and ggplot2.continuous.fill. These can be set to "gradient" (the default) or "viridis" (@karawoo).

  • New scale_colour_viridis_c()/scale_fill_viridis_c() (continuous) and scale_colour_viridis_d()/scale_fill_viridis_d() (discrete) make it easy to use Viridis colour scales (@karawoo, #1526).

  • Guides for geom_text() now accept custom labels with guide_legend(override.aes = list(label = "foo")) (@brianwdavis, #2458).

Margins

  • Strips gain margins on all sides by default. This means that to fully justify text to the edge of a strip, you will need to also set the margins to 0 (@karawoo).

  • Rotated strip labels now correctly understand hjust and vjust parameters at all angles (@karawoo).

  • Strip labels now understand justification relative to the direction of the text, meaning that in y facets, the strip text can be placed at either end of the strip using hjust (@karawoo).

  • Legend titles and labels get a little extra space around them, which prevents legend titles from overlapping the legend at large font sizes (@karawoo, #1881).

Extension points

  • New autolayer() S3 generic (@mitchelloharawild, #1974). This is similar to autoplot() but produces layers rather than complete plots.

  • Custom objects can now be added using + if a ggplot_add method has been defined for the class of the object (@thomasp85).

  • Theme elements can now be subclassed. Add a merge_element method to control how properties are inherited from the parent element. Add an element_grob method to define how elements are rendered into grobs (@thomasp85, #1981).

  • Coords have gained new extension mechanisms.

    If you have an existing coord extension, you will need to revise the specification of the train() method. It is now called setup_panel_params() (better reflecting what it actually does) and now has arguments scale_x, and scale_y (the x and y scales respectively) and param, a list of plot specific parameters generated by setup_params().

    What was formerly called scale_details (in coords), panel_ranges (in layout) and panel_scales (in geoms) are now consistently called panel_params (#1311). These are parameters of the coord that vary from panel to panel.

  • ggplot_build() and ggplot_gtable() are now generics, so ggplot-subclasses can define additional behavior during the build stage.

  • guide_train(), guide_merge(), guide_geom(), and guide_gengrob() are now exported as they are needed if you want to design your own guide. They are not currently documented; use at your own risk (#2528).

  • scale_type() generic is now exported and documented. Use this if you want to extend ggplot2 to work with a new type of vector.

Minor bug fixes and improvements

Faceting

  • facet_grid() gives a more informative error message if you try to use a variable in both rows and cols (#1928).

  • facet_grid() and facet_wrap() both give better error messages if you attempt to use an unsupported coord with free scales (#2049).

  • label_parsed() works once again (#2279).

  • You can now style the background of horizontal and vertical strips independently with strip.background.x and strip.background.y theme settings (#2249).

Scales

Layers

Coords

  • Clipping to the plot panel is now configurable, through a clip argument to coordinate systems, e.g. coord_cartesian(clip = "off") (@clauswilke, #2536).

  • Like scales, coordinate systems now give you a message when you’re replacing an existing coordinate system (#2264).

  • coord_polar() now draws secondary axis ticks and labels (@dylan-stark, #2072), and can draw the radius axis on the right (@thomasp85, #2005).

  • coord_trans() now generates a warning when a transformation generates non-finite values (@foo-bar-baz-qux, #2147).

Themes

  • Complete themes now always override all elements of the default theme (@has2k1, #2058, #2079).

  • Themes now set default grid colour in panel.grid rather than individually in panel.grid.major and panel.grid.minor individually. This makes it slightly easier to customise the theme (#2352).

  • Fixed bug when setting strips to element_blank() (@thomasp85).

  • Axes positioned on the top and to the right can now customize their ticks and lines separately (@thomasp85, #1899).

  • Built-in themes gain parameters base_line_size and base_rect_size which control the default sizes of line and rectangle elements (@karawoo, #2176).

  • Default themes use rel() to set line widths (@baptiste).

  • Themes were tweaked for visual consistency and more graceful behavior when changing the base font size. All absolute heights or widths were replaced with heights or widths that are proportional to the base font size. One relative font size was eliminated (@clauswilke).

  • The height of descenders is now calculated solely on font metrics and doesn’t change with the specific letters in the string. This fixes minor alignment issues with plot titles, subtitles, and legend titles (#2288, @clauswilke).

Guides

Other

  • fortify() gains a method for tbls (@karawoo, #2218).

  • ggplot gains a method for grouped_dfs that adds a .group variable, which computes a unique value for each group. Use it with aes(group = .group) (#2351).

  • ggproto() produces objects with class c("ggproto", "gg"), allowing for a more informative error message when adding layers, scales, or other ggproto objects (@jrnold, #2056).

  • ggsave()’s DPI argument now supports 3 string options: “retina” (320 DPI), “print” (300 DPI), and “screen” (72 DPI) (@foo-bar-baz-qux, #2156). ggsave() now uses full argument names to avoid partial match warnings (#2355), and correctly restores the previous graphics device when several graphics devices are open (#2363).

  • print.ggplot() now returns the original ggplot object, instead of the output from ggplot_build(). Also, the object returned from ggplot_build() now has the class "ggplot_built" (#2034).

  • map_data() now works even when purrr is loaded (tidyverse#66).

  • New functions summarise_layout(), summarise_coord(), and summarise_layers() summarise the layout, coordinate systems, and layers of a built ggplot object (#2034, @wch). This provides a tested API that (e.g.) shiny can depend on.

  • Updated startup messages reflect new resources (#2410, @mine-cetinkaya-rundel).

ggplot2 2.2.1

CRAN release: 2016-12-30

  • Fix usage of structure(NULL) for R-devel compatibility (#1968).

ggplot2 2.2.0

CRAN release: 2016-11-11

Major new features

Subtitle and caption

Thanks to @hrbrmstr plots now have subtitles and captions, which can be set with the subtitle and caption arguments to ggtitle() and labs(). You can control their appearance with the theme settings plot.caption and plot.subtitle. The main plot title is now left-aligned to better work better with a subtitle. The caption is right-aligned (@hrbrmstr).

Stacking

position_stack() and position_fill() now sort the stacking order to match grouping order. This allows you to control the order through grouping, and ensures that the default legend matches the plot (#1552, #1593). If you want the opposite order (useful if you have horizontal bars and horizontal legend), you can request reverse stacking by using position = position_stack(reverse = TRUE) (#1837).

position_stack() and position_fill() now accepts negative values which will create stacks extending below the x-axis (#1691).

position_stack() and position_fill() gain a vjust argument which makes it easy to (e.g.) display labels in the middle of stacked bars (#1821).

Layers

geom_col() was added to complement geom_bar() (@hrbrmstr). It uses stat="identity" by default, making the y aesthetic mandatory. It does not support any other stat_() and does not provide fallback support for the binwidth parameter. Examples and references in other functions were updated to demonstrate geom_col() usage.

When creating a layer, ggplot2 will warn if you use an unknown aesthetic or an unknown parameter. Compared to the previous version, this is stricter for aesthetics (previously there was no message), and less strict for parameters (previously this threw an error) (#1585).

Facetting

The facet system, as well as the internal panel class, has been rewritten in ggproto. Facets are now extendable in the same manner as geoms and stats, as described in vignette("extending-ggplot2").

We have also added the following new features.

  • facet_grid() and facet_wrap() now allow expressions in their faceting formulas (@DanRuderman, #1596).

  • When facet_wrap() results in an uneven number of panels, axes will now be drawn underneath the hanging panels (fixes #1607)

  • Strips can now be freely positioned in facet_wrap() using the strip.position argument (deprecates switch).

  • The relative order of panel, strip, and axis can now be controlled with the theme setting strip.placement that takes either inside (strip between panel and axis) or outside (strip after axis).

  • The theme option panel.margin has been deprecated in favour of panel.spacing to more clearly communicate intent.

Extensions

Unfortunately there was a major oversight in the construction of ggproto which lead to extensions capturing the super object at package build time, instead of at package run time (#1826). This problem has been fixed, but requires re-installation of all extension packages.

Scales

  • The position of x and y axes can now be changed using the position argument in scale_x_*and scale_y_* which can take top and bottom, and left and right respectively. The themes of top and right axes can be modified using the .top and .right modifiers to axis.text.* and axis.title.*.

Continuous scales

  • scale_x_continuous() and scale_y_continuous() can now display a secondary axis that is a one-to-one transformation of the primary axis (e.g. degrees Celcius to degrees Fahrenheit). The secondary axis will be positioned opposite to the primary axis and can be controlled with the sec.axis argument to the scale constructor.

  • Scales worry less about having breaks. If no breaks can be computed, the plot will work instead of throwing an uninformative error (#791). This is particularly helpful when you have facets with free scales, and not all panels contain data.

  • Scales now warn when transformation introduces infinite values (#1696).

Date time

  • scale_*_datetime() now supports time zones. It will use the timezone attached to the variable by default, but can be overridden with the timezone argument.

  • New scale_x_time() and scale_y_time() generate reasonable default breaks and labels for hms vectors (#1752).

Discrete scales

The treatment of missing values by discrete scales has been thoroughly overhauled (#1584). The underlying principle is that we can naturally represent missing values on discrete variables (by treating just like another level), so by default we should.

This principle applies to:

  • character vectors
  • factors with implicit NA
  • factors with explicit NA

And to all scales (both position and non-position.)

Compared to the previous version of ggplot2, there are three main changes:

  1. scale_x_discrete() and scale_y_discrete() always show discrete NA, regardless of their source

  2. If present, NAs are shown in discrete legends.

  3. All discrete scales gain a na.translate argument that allows you to control whether NAs are translated to something that can be visualised, or should be left as missing. Note that if you don’t translate (i.e.  na.translate = FALSE) the missing values will passed on to the layer, which will warning that it’s dropping missing values. To suppress the warnings, you’ll also need to add na.rm = TRUE to the layer call.

There were also a number of other smaller changes

  • Correctly use scale expansion factors.
  • Don’t preserve space for dropped levels (#1638).
  • Only issue one warning when when asking for too many levels (#1674).
  • Unicode labels work better on Windows (#1827).
  • Warn when used with only continuous data (#1589)

Themes

  • The theme() constructor now has named arguments rather than ellipses. This should make autocomplete substantially more useful. The documentation (including examples) has been considerably improved.

  • Built-in themes are more visually homogeneous, and match theme_grey better. (@jiho, #1679)

  • When computing the height of titles, ggplot2 now includes the height of the descenders (i.e. the bits of g and y that hang beneath the baseline). This improves the margins around titles, particularly the y axis label (#1712). I have also very slightly increased the inner margins of axis titles, and removed the outer margins.

  • Theme element inheritance is now easier to work with as modification now overrides default element_blank elements (#1555, #1557, #1565, #1567)

  • Horizontal legends (i.e. legends on the top or bottom) are horizontally aligned by default (#1842). Use legend.box = "vertical" to switch back to the previous behaviour.

  • element_line() now takes an arrow argument to specify arrows at the end of lines (#1740)

There were a number of tweaks to the theme elements that control legends:

  • legend.justification now controls appearance will plotting the legend outside of the plot area. For example, you can use theme(legend.justification = "top") to make the legend align with the top of the plot.

  • panel.margin and legend.margin have been renamed to panel.spacing and legend.spacing respectively, to better communicate intent (they only affect spacing between legends and panels, not the margins around them)

  • legend.margin now controls margin around individual legends.

  • New legend.box.background, legend.box.spacing, and legend.box.margin control the background, spacing, and margin of the legend box (the region that contains all legends).

Bug fixes and minor improvements

  • ggplot2 now imports tibble. This ensures that all built-in datasets print compactly even if you haven’t explicitly loaded tibble or dplyr (#1677).

  • Class of aesthetic mapping is preserved when adding aes() objects (#1624).

  • +.gg now works for lists that include data frames.

  • annotation_x() now works in the absense of global data (#1655)

  • geom_*(show.legend = FALSE) now works for guide_colorbar.

  • geom_boxplot() gains new outlier.alpha (@jonathan-g) and outlier.fill (@schloerke, #1787) parameters to control the alpha/fill of outlier points independently of the alpha of the boxes.

  • position_jitter() (and hence geom_jitter()) now correctly computes the jitter width/jitter when supplied by the user (#1775, @has2k1).

  • geom_contour() more clearly describes what inputs it needs (#1577).

  • geom_curve() respects the lineend parameter (#1852).

  • geom_histogram() and stat_bin() understand the breaks parameter once more. (#1665). The floating point adjustment for histogram bins is now actually used - it was previously inadvertently ignored (#1651).

  • geom_violin() no longer transforms quantile lines with the alpha aesthetic (@mnbram, #1714). It no longer errors when quantiles are requested but data have zero range (#1687). When trim = FALSE it once again has a nice range that allows the density to reach zero (by extending the range 3 bandwidths to either side of the data) (#1700).

  • geom_dotplot() works better when faceting and binning on the y-axis. (#1618, @has2k1).

  • geom_hexbin() once again supports ..density.. (@mikebirdgeneau, #1688).

  • geom_step() gives useful warning if only one data point in layer (#1645).

  • layer() gains new check.aes and check.param arguments. These allow geom/stat authors to optional suppress checks for known aesthetics/parameters. Currently this is used only in geom_blank() which powers expand_limits() (#1795).

  • All stat_*() display a better error message when required aesthetics are missing.

  • stat_bin() and stat_summary_hex() now accept length 1 binwidth (#1610)

  • stat_density() gains new argument n, which is passed to underlying function stats::density (“number of equally spaced points at which the density is to be estimated”). (@hbuschme)

  • stat_binhex() now again returns count rather than value (#1747)

  • stat_ecdf() respects pad argument (#1646).

  • stat_smooth() once again informs you about the method it has chosen. It also correctly calculates the size of the largest group within facets.

  • x and y scales are now symmetric regarding the list of aesthetics they accept: xmin_final, xmax_final, xlower, xmiddle and xupper are now valid x aesthetics.

  • Scale extensions can now override the make_title and make_sec_title methods to let the scale modify the axis/legend titles.

  • The random stream is now reset after calling .onAttach() (#2409).

ggplot2 2.1.0

CRAN release: 2016-03-01

New features

  • When mapping an aesthetic to a constant (e.g.  geom_smooth(aes(colour = "loess")))), the default guide title is the name of the aesthetic (i.e. “colour”), not the value (i.e. “loess”) (#1431).

  • layer() now accepts a function as the data argument. The function will be applied to the data passed to the ggplot() function and must return a data.frame (#1527, @thomasp85). This is a more general version of the deprecated subset argument.

  • theme_update() now uses the + operator instead of %+replace%, so that unspecified values will no longer be NULLed out. theme_replace() preserves the old behaviour if desired (@oneillkza, #1519).

  • stat_bin() has been overhauled to use the same algorithm as ggvis, which has been considerably improved thanks to the advice of Randy Prium (@rpruim). This includes:

    • Better arguments and a better algorithm for determining the origin. You can now specify either boundary or the center of a bin. origin has been deprecated in favour of these arguments.

    • drop is deprecated in favour of pad, which adds extra 0-count bins at either end (needed for frequency polygons). geom_histogram() defaults to pad = FALSE which considerably improves the default limits for the histogram, especially when the bins are big (#1477).

    • The default algorithm does a (somewhat) better job at picking nice widths and origins across a wider range of input data.

    • bins = n now gives a histogram with n bins, not n + 1 (#1487).

Bug fixes

  • All \donttest{} examples run.

  • All geom_() and stat_() functions now have consistent argument order: data + mapping, then geom/stat/position, then ..., then specific arguments, then arguments common to all layers (#1305). This may break code if you were previously relying on partial name matching, but in the long-term should make ggplot2 easier to use. In particular, you can now set the n parameter in geom_density2d() without it partially matching na.rm (#1485).

  • For geoms with both colour and fill, alpha once again only affects fill (Reverts #1371, #1523). This was causing problems for people.

  • facet_wrap()/facet_grid() works with multiple empty panels of data (#1445).

  • facet_wrap() correctly swaps nrow and ncol when faceting vertically (#1417).

  • ggsave("x.svg") now uses svglite to produce the svg (#1432).

  • geom_boxplot() now understands outlier.color (#1455).

  • geom_path() knows that “solid” (not just 1) represents a solid line (#1534).

  • geom_ribbon() preserves missing values so they correctly generate a gap in the ribbon (#1549).

  • geom_tile() once again accepts width and height parameters (#1513). It uses draw_key_polygon() for better a legend, including a coloured outline (#1484).

  • layer() now automatically adds a na.rm parameter if none is explicitly supplied.

  • position_jitterdodge() now works on all possible dodge aesthetics, e.g. color, linetype etc. instead of only based on fill (@bleutner)

  • position = "nudge" now works (although it doesn’t do anything useful) (#1428).

  • The default scale for columns of class “AsIs” is now “identity” (#1518).

  • scale_*_discrete() has better defaults when used with purely continuous data (#1542).

  • scale_size() warns when used with categorical data.

  • scale_size(), scale_colour(), and scale_fill() gain date and date-time variants (#1526).

  • stat_bin_hex() and stat_bin_summary() now use the same underlying algorithm so results are consistent (#1383). stat_bin_hex() now accepts a weight aesthetic. To be consistent with related stats, the output variable from stat_bin_hex() is now value instead of count.

  • stat_density() gains a bw parameter which makes it easy to get consistent smoothing between facets (@jiho)

  • stat-density-2d() no longer ignores the h parameter, and now accepts bins and binwidth parameters to control the number of contours (#1448, @has2k1).

  • stat_ecdf() does a better job of adding padding to -Inf/Inf, and gains an argument pad to suppress the padding if not needed (#1467).

  • stat_function() gains an xlim parameter (#1528). It once again works with discrete x values (#1509).

  • stat_summary() preserves sorted x order which avoids artefacts when display results with geom_smooth() (#1520).

  • All elements should now inherit correctly for all themes except theme_void(). (@Katiedaisey, #1555)

  • theme_void() was completely void of text but facets and legends still need labels. They are now visible (@jiho).

  • You can once again set legend key and height width to unit arithmetic objects (like 2 * unit(1, "cm")) (#1437).

  • Eliminate spurious warning if you have a layer with no data and no aesthetics (#1451).

  • Removed a superfluous comma in theme-defaults.r code (@jschoeley)

  • Fixed a compatibility issue with ggproto and R versions prior to 3.1.2. (#1444)

  • Fixed issue where coord_map() fails when given an explicit parameters argument (@tdmcarthur, #1729)

  • Fixed issue where geom_errorbarh() had a required x aesthetic (#1933)

ggplot2 2.0.0

CRAN release: 2015-12-18

Major changes

  • ggplot no longer throws an error if your plot has no layers. Instead it automatically adds geom_blank() (#1246).

  • New cut_width() is a convenient replacement for the verbose plyr::round_any(), with the additional benefit of offering finer control.

  • New geom_count() is a convenient alias to stat_sum(). Use it when you have overlapping points on a scatterplot. stat_sum() now defaults to using counts instead of proportions.

  • New geom_curve() adds curved lines, with a similar specification to geom_segment() (@veraanadi, #1088).

  • Date and datetime scales now have date_breaks, date_minor_breaks and date_labels arguments so that you never need to use the long scales::date_breaks() or scales::date_format().

  • geom_bar() now has it’s own stat, distinct from stat_bin() which was also used by geom_histogram(). geom_bar() now uses stat_count() which counts values at each distinct value of x (i.e. it does not bin the data first). This can be useful when you want to show exactly which values are used in a continuous variable.

  • geom_point() gains a stroke aesthetic which controls the border width of shapes 21-25 (#1133, @SeySayux). size and stroke are additive so a point with size = 5 and stroke = 5 will have a diameter of 10mm. (#1142)

  • New position_nudge() allows you to slightly offset labels (or other geoms) from their corresponding points (#1109).

  • scale_size() now maps values to area, not radius. Use scale_radius() if you want the old behaviour (not recommended, except perhaps for lines).

  • New stat_summary_bin() works like stat_summary() but on binned data. It’s a generalisation of stat_bin() that can compute any aggregate, not just counts (#1274). Both default to mean_se() if no aggregation functions are supplied (#1386).

  • Layers are now much stricter about their arguments - you will get an error if you’ve supplied an argument that isn’t an aesthetic or a parameter. This is likely to cause some short-term pain but in the long-term it will make it much easier to spot spelling mistakes and other errors (#1293).

    This change does break a handful of geoms/stats that used ... to pass additional arguments on to the underlying computation. Now geom_smooth()/stat_smooth() and geom_quantile()/stat_quantile() use method.args instead (#1245, #1289); and stat_summary() (#1242), stat_summary_hex(), and stat_summary2d() use fun.args.

Extensibility

There is now an official mechanism for defining Stats, Geoms, and Positions in other packages. See vignette("extending-ggplot2") for details.

  • All Geoms, Stats and Positions are now exported, so you can inherit from them when making your own objects (#989).

  • ggplot2 no longer uses proto or reference classes. Instead, we now use ggproto, a new OO system designed specifically for ggplot2. Unlike proto and RC, ggproto supports clean cross-package inheritance. Creating a new OO system isn’t usually the right way to solve a problem, but I’m pretty sure it was necessary here. Read more about it in the vignette.

  • aes_() replaces aes_q(). It also supports formulas, so the most concise SE version of aes(carat, price) is now aes_(~carat, ~price). You may want to use this form in packages, as it will avoid spurious R CMD check warnings about undefined global variables.

Text

  • geom_text() has been overhauled to make labelling your data a little easier. It:

    • nudge_x and nudge_y arguments let you offset labels from their corresponding points (#1120).

    • check_overlap = TRUE provides a simple way to avoid overplotting of labels: labels that would otherwise overlap are omitted (#1039).

    • hjust and vjust can now be character vectors: “left”, “center”, “right”, “bottom”, “middle”, “top”. New options include “inward” and “outward” which align text towards and away from the center of the plot respectively.

  • geom_label() works like geom_text() but draws a rounded rectangle underneath each label (#1039). This is useful when you want to label plots that are dense with data.

Deprecated features

  • The little used aes_auto() has been deprecated.

  • aes_q() has been replaced with aes_() to be consistent with SE versions of NSE functions in other packages.

  • The order aesthetic is officially deprecated. It never really worked, and was poorly documented.

  • The stat and position arguments to qplot() have been deprecated. qplot() is designed for quick plots - if you need to specify position or stat, use ggplot() instead.

  • The theme setting axis.ticks.margin has been deprecated: now use the margin property of axis.text.

  • stat_abline(), stat_hline() and stat_vline() have been removed: these were never suitable for use other than with geom_abline() etc and were not documented.

  • show_guide has been renamed to show.legend: this more accurately reflects what it does (controls appearance of layer in legend), and uses the same convention as other ggplot2 arguments (i.e. a . between names). (Yes, I know that’s inconsistent with function names with use _, but it’s too late to change now.)

A number of geoms have been renamed to be internally consistent:

All defunct functions have been removed.

Default appearance

  • The default theme_grey() background colour has been changed from “grey90” to “grey92”: this makes the background a little less visually prominent.

  • Labels and titles have been tweaked for readability:

    • Axes labels are darker.

    • Legend and axis titles are given the same visual treatment.

    • The default font size dropped from 12 to 11. You might be surprised that I’ve made the default text size smaller as it was already hard for many people to read. It turns out there was a bug in RStudio (fixed in 0.99.724), that shrunk the text of all grid based graphics. Once that was resolved the defaults seemed too big to my eyes.

    • More spacing between titles and borders.

    • Default margins scale with the theme font size, so the appearance at larger font sizes should be considerably improved (#1228).

  • alpha now affects both fill and colour aesthetics (#1371).

  • element_text() gains a margins argument which allows you to add additional padding around text elements. To help see what’s going on use debug = TRUE to display the text region and anchors.

  • The default font size in geom_text() has been decreased from 5mm (14 pts) to 3.8 mm (11 pts) to match the new default theme sizes.

  • A diagonal line is no longer drawn on bar and rectangle legends. Instead, the border has been tweaked to be more visible, and more closely match the size of line drawn on the plot.

  • geom_pointrange() and geom_linerange() get vertical (not horizontal) lines in the legend (#1389).

  • The default line size for geom_smooth() has been increased from 0.5 to 1 to make it easier to see when overlaid on data.

  • geom_bar() and geom_rect() use a slightly paler shade of grey so they aren’t so visually heavy.

  • geom_boxplot() now colours outliers the same way as the boxes.

  • geom_point() now uses shape 19 instead of 16. This looks much better on the default Linux graphics device. (It’s very slightly smaller than the old point, but it shouldn’t affect any graphics significantly)

  • Sizes in ggplot2 are measured in mm. Previously they were converted to pts (for use in grid) by multiplying by 72 / 25.4. However, grid uses printer’s points, not Adobe (big pts), so sizes are now correctly multiplied by 72.27 / 25.4. This is unlikely to noticeably affect display, but it’s technically correct (https://youtu.be/hou0lU8WMgo).

  • The default legend will now allocate multiple rows (if vertical) or columns (if horizontal) in order to make a legend that is more likely to fit on the screen. You can override with the nrow/ncol arguments to guide_legend()

    
    p <- ggplot(mpg, aes(displ,hwy, colour = model)) + geom_point()
    p
    p + theme(legend.position = "bottom")
    # Previous behaviour
    p + guides(colour = guide_legend(ncol = 1))

New and updated themes

  • New theme_void() is completely empty. It’s useful for plots with non- standard coordinates or for drawings (@jiho, #976).

  • New theme_dark() has a dark background designed to make colours pop out (@jiho, #1018)

  • theme_minimal() became slightly more minimal by removing the axis ticks: labels now line up directly beneath grid lines (@tomschloss, #1084)

  • New theme setting panel.ontop (logical) make it possible to place background elements (i.e., gridlines) on top of data. Best used with transparent panel.background (@noamross. #551).

Labelling

The facet labelling system was updated with many new features and a more flexible interface (@lionel-). It now works consistently across grid and wrap facets. The most important user visible changes are:

  • facet_wrap() gains a labeller option (#25).

  • facet_grid() and facet_wrap() gain a switch argument to display the facet titles near the axes. When switched, the labels become axes subtitles. switch can be set to “x”, “y” or “both” (the latter only for grids) to control which margin is switched.

The labellers (such as label_value() or label_both()) also get some new features:

  • They now offer the multi_line argument to control whether to display composite facets (those specified as ~var1 + var2) on one or multiple lines.

  • In label_bquote() you now refer directly to the names of variables. With this change, you can create math expressions that depend on more than one variable. This math expression can be specified either for the rows or the columns and you can also provide different expressions to each margin.

    As a consequence of these changes, referring to x in backquoted expressions is deprecated.

  • Similarly to label_bquote(), labeller() now take .rows and .cols arguments. In addition, it also takes .default. labeller() is useful to customise how particular variables are labelled. The three additional arguments specify how to label the variables are not specifically mentioned, respectively for rows, columns or both. This makes it especially easy to set up a project-wide labeller dispatcher that can be reused across all your plots. See the documentation for an example.

  • The new labeller label_context() adapts to the number of factors facetted over. With a single factor, it displays only the values, just as before. But with multiple factors in a composite margin (e.g. with ~cyl + am), the labels are passed over to label_both(). This way the variables names are displayed with the values to help identifying them.

On the programming side, the labeller API has been rewritten in order to offer more control when faceting over multiple factors (e.g. with formulae such as ~cyl + am). This also means that if you have written custom labellers, you will need to update them for this version of ggplot.

  • Previously, a labeller function would take variable and value arguments and return a character vector. Now, they take a data frame of character vectors and return a list. The input data frame has one column per factor facetted over and each column in the returned list becomes one line in the strip label. See documentation for more details.

  • The labels received by a labeller now contain metadata: their margin (in the “type” attribute) and whether they come from a wrap or a grid facet (in the “facet” attribute).

  • Note that the new as_labeller() function operator provides an easy way to transform an existing function to a labeller function. The existing function just needs to take and return a character vector.

Documentation

  • Improved documentation for aes(), layer() and much much more.

  • I’ve tried to reduce the use of ... so that you can see all the documentation in one place rather than having to integrate multiple pages. In some cases this has involved adding additional arguments to geoms to make it more clear what you can do:

  • Use of qplot() in examples has been minimised (#1123, @hrbrmstr). This is inline with the 2nd edition of the ggplot2 box, which minimises the use of qplot() in favour of ggplot().

  • Tightly linked geoms and stats (e.g. geom_boxplot() and stat_boxplot()) are now documented in the same file so you can see all the arguments in one place. Variations of the same idea (e.g. geom_path(), geom_line(), and geom_step()) are also documented together.

  • It’s now obvious that you can set the binwidth parameter for stat_bin_hex(), stat_summary_hex(), stat_bin_2d(), and stat_summary_2d().

  • The internals of positions have been cleaned up considerably. You’re unlikely to notice any external changes, although the documentation should be a little less confusing since positions now don’t list parameters they never use.

Data

  • All datasets have class tbl_df so if you also use dplyr, you get a better print method.

  • economics has been brought up to date to 2015-04-01.

  • New economics_long is the economics data in long form.

  • New txhousing dataset containing information about the Texas housing market. Useful for examples that need multiple time series, and for demonstrating model+vis methods.

  • New luv_colours dataset which contains the locations of all built-in colors() in Luv space.

  • movies has been moved into its own package, ggplot2movies, because it was large and not terribly useful. If you’ve used the movies dataset, you’ll now need to explicitly load the package with library(ggplot2movies).

Bug fixes and minor improvements

  • All partially matched arguments and $ have been been replaced with full matches (@jimhester, #1134).

  • ggplot2 now exports alpha() from the scales package (#1107), and arrow() and unit() from grid (#1225). This means you don’t need attach scales/grid or do scales::/grid:: for these commonly used functions.

  • aes_string() now only parses character inputs. This fixes bugs when using it with numbers and non default OutDec settings (#1045).

  • annotation_custom() automatically adds a unique id to each grob name, making it easier to plot multiple grobs with the same name (e.g. grobs of ggplot2 graphics) in the same plot (#1256).

  • borders() now accepts xlim and ylim arguments for specifying the geographical region of interest (@markpayneatwork, #1392).

  • coord_cartesian() applies the same expansion factor to limits as for scales. You can suppress with expand = FALSE (#1207).

  • coord_trans() now works when breaks are suppressed (#1422).

  • cut_number() gives error message if the number of requested bins can be created because there are two few unique values (#1046).

  • Character labels in facet_grid() are no longer (incorrectly) coerced into factors. This caused problems with custom label functions (#1070).

  • facet_wrap() and facet_grid() now allow you to use non-standard variable names by surrounding them with backticks (#1067).

  • facet_wrap() more carefully checks its nrow and ncol arguments to ensure that they’re specified correctly (@richierocks, #962)

  • facet_wrap() gains a dir argument to control the direction the panels are wrapped in. The default is “h” for horizontal. Use “v” for vertical layout (#1260).

  • geom_abline(), geom_hline() and geom_vline() have been rewritten to have simpler behaviour and be more consistent:

  • geom_bin2d() will now let you specify one dimension’s breaks exactly, without touching the other dimension’s default breaks at all (#1126).

  • geom_crossbar() sets grouping correctly so you can display multiple crossbars on one plot. It also makes the default fatten argument a little bigger to make the middle line more obvious (#1125).

  • geom_histogram() and geom_smooth() now only inform you about the default values once per layer, rather than once per panel (#1220).

  • geom_pointrange() gains fatten argument so you can control the size of the point relative to the size of the line.

  • geom_segment() annotations were not transforming with scales (@BrianDiggs, #859).

  • geom_smooth() is no longer so chatty. If you want to know what the default smoothing method is, look it up in the documentation! (#1247)

  • geom_violin() now has the ability to draw quantile lines (@DanRuderman).

  • ggplot() now captures the parent frame to use for evaluation, rather than always defaulting to the global environment. This should make ggplot more suitable to use in more situations (e.g. with knitr)

  • ggsave() has been simplified a little to make it easier to maintain. It no longer checks that you’re printing a ggplot2 object (so now also works with any grid grob) (#970), and always requires a filename. Parameter device now supports character argument to specify which supported device to use (‘pdf’, ‘png’, ‘jpeg’, etc.), for when it cannot be correctly inferred from the file extension (for example when a temporary filename is supplied server side in shiny apps) (@sebkopf, #939). It no longer opens a graphics device if one isn’t already open - this is annoying when you’re running from a script (#1326).

  • guide_colorbar() creates correct legend if only one color (@krlmlr, #943).

  • guide_colorbar() no longer fails when the legend is empty - previously this often masked misspecifications elsewhere in the plot (#967).

  • New layer_data() function extracts the data used for plotting for a given layer. It’s mostly useful for testing.

  • User supplied minor_breaks can now be supplied on the same scale as the data, and will be automatically transformed with by scale (#1385).

  • You can now suppress the appearance of an axis/legend title (and the space that would allocated for it) with NULL in the scale_ function. To use the default label, use waiver() (#1145).

  • Position adjustments no longer warn about potentially varying ranges because the problem rarely occurs in practice and there are currently a lot of false positives since I don’t understand exactly what FP criteria I should be testing.

  • scale_fill_grey() now uses red for missing values. This matches scale_colour_grey() and makes it obvious where missing values lie. Override with na.value.

  • scale_*_gradient2() defaults to using Lab colour space.

  • scale_*_gradientn() now allows colours or colors (#1290)

  • scale_y_continuous() now also transforms the lower, middle and upper aesthetics used by geom_boxplot(): this only affects geom_boxplot(stat = "identity") (#1020).

  • Legends no longer inherit aesthetics if inherit.aes is FALSE (#1267).

  • lims() makes it easy to set the limits of any axis (#1138).

  • labels = NULL now works with guide_legend() and guide_colorbar(). (#1175, #1183).

  • override.aes now works with American aesthetic spelling, e.g. color

  • Scales no longer round data points to improve performance of colour palettes. Instead the scales package now uses a much faster colour interpolation algorithm (#1022).

  • scale_*_brewer() and scale_*_distiller() add new direction argument of scales::brewer_pal, making it easier to change the order of colours (@jiho, #1139).

  • scale_x_date() now clips dates outside the limits in the same way as scale_x_continuous() (#1090).

  • stat_bin() gains bins arguments, which denotes the number of bins. Now you can set bins=100 instead of binwidth=0.5. Note that breaks or binwidth will override it (@tmshn, #1158, #102).

  • stat_boxplot() warns if a continuous variable is used for the x aesthetic without also supplying a group aesthetic (#992, @krlmlr).

  • stat_summary_2d() and stat_bin_2d() now share exactly the same code for determining breaks from bins, binwidth, and origin.

  • stat_summary_2d() and stat_bin_2d() now output in tile/raster compatible form instead of rect compatible form.

  • Automatically computed breaks do not lead to an error for transformations like “probit” where the inverse can map to infinity (#871, @krlmlr)

  • stat_function() now always evaluates the function on the original scale. Previously it computed the function on transformed scales, giving incorrect values (@BrianDiggs, #1011).

  • strip_dots works with anonymous functions within calculated aesthetics (e.g. aes(sapply(..density.., function(x) mean(x)))) (#1154, @NikNakk)

  • theme() gains validate = FALSE parameter to turn off validation, and hence store arbitrary additional data in the themes. (@tdhock, #1121)

  • Improved the calculation of segments needed to draw the curve representing a line when plotted in polar coordinates. In some cases, the last segment of a multi-segment line was not drawn (@BrianDiggs, #952)