Create a ggplot in a Single Call

plot2() creates a complete ggplot in one call. It takes the place of the usual combination of ggplot(), aes(), a geom_*() function, scales, labels and a theme, and of the data preparation that often precedes it with group_by(), count(), mutate() or summarise(): calculations can be given in the arguments directly, such as y = n() or x = format(date, "%Y").

Besides data frames, plot2() accepts vectors, matrices, formulas, 'sf' objects and other classes; see plot2-methods.

Usage

plot2(
  .data,
  x = NULL,
  y = NULL,
  category = NULL,
  facet = NULL,
  type = NULL,
  x.title = TRUE,
  y.title = TRUE,
  category.title = NULL,
  title = NULL,
  subtitle = NULL,
  caption = NULL,
  tag = NULL,
  title.linelength = 60,
  title.colour = getOption("plot2.colour_font_primary", "black"),
  subtitle.linelength = 60,
  subtitle.colour = getOption("plot2.colour_font_secondary", "grey35"),
  na.replace = "",
  na.rm = FALSE,
  facet.position = "top",
  facet.colour_fill = NULL,
  facet.bold = TRUE,
  facet.italic = FALSE,
  facet.size = 10,
  facet.margin = 8,
  facet.repeat_labels_x = TRUE,
  facet.repeat_labels_y = NULL,
  facet.fixed_y = NULL,
  facet.fixed_x = NULL,
  facet.drop = FALSE,
  facet.nrow = NULL,
  facet.relative = FALSE,
  x.date_breaks = NULL,
  x.date_labels = NULL,
  x.date_remove_years = NULL,
  category.focus = NULL,
  colour = getOption("plot2.colour", "ggplot2"),
  colour_fill = NULL,
  colour_transparency = 0,
  x.label_angle = 0,
  x.label_align = NULL,
  x.label_italic = FALSE,
  x.remove = FALSE,
  x.position = "bottom",
  x.max_items = NULL,
  x.max_txt = "(rest, x%n)",
  category.max_items = Inf,
  category.max_txt = "(rest, x%n)",
  facet.max_items = Inf,
  facet.max_txt = "(rest, x%n)",
  x.breaks = NULL,
  x.n_breaks = NULL,
  x.transform = "identity",
  x.expand = NULL,
  x.limits = NULL,
  x.labels = NULL,
  x.character = NULL,
  x.drop = FALSE,
  x.zoom = NULL,
  y.remove = FALSE,
  y.24h = FALSE,
  y.age = FALSE,
  y.scientific = NULL,
  y.percent = NULL,
  y.percent_break = NULL,
  y.breaks = NULL,
  y.n_breaks = NULL,
  y.limits = NULL,
  y.labels = NULL,
  y.expand = NULL,
  y.transform = "identity",
  y.position = "left",
  y.zoom = FALSE,
  y_secondary = NULL,
  y_secondary.type = NULL,
  y_secondary.title = TRUE,
  y_secondary.colour = NULL,
  y_secondary.colour_fill = NULL,
  y_secondary.scientific = NULL,
  y_secondary.percent = FALSE,
  y_secondary.labels = NULL,
  category.type = "colour",
  category.labels = NULL,
  category.percent = FALSE,
  category.breaks = NULL,
  category.limits = NULL,
  category.expand = 0,
  category.midpoint = NULL,
  category.transform = "identity",
  category.date_breaks = NULL,
  category.date_labels = NULL,
  category.character = NULL,
  x.sort = NULL,
  category.sort = TRUE,
  facet.sort = TRUE,
  x.complete = NULL,
  category.complete = NULL,
  facet.complete = NULL,
  datalabels = NULL,
  datalabels.round = ifelse(isTRUE(y.percent), 2, 1),
  datalabels.format = "%n",
  datalabels.colour = "grey25",
  datalabels.colour_fill = NULL,
  datalabels.size = (3 * text_factor),
  datalabels.angle = 0,
  datalabels.lineheight = 1,
  datalabels.percent_by = "x",
  decimal.mark = dec_mark(),
  big.mark = big_mark(),
  summarise_function = NULL,
  stacked = FALSE,
  stacked_fill = FALSE,
  horizontal = FALSE,
  reverse = horizontal,
  smooth = NULL,
  smooth.method = NULL,
  smooth.formula = NULL,
  smooth.se = TRUE,
  smooth.level = 0.95,
  smooth.alpha = NULL,
  smooth.linewidth = NULL,
  smooth.linetype = NULL,
  smooth.colour = NULL,
  size = NULL,
  linetype = 1,
  linewidth = NULL,
  binwidth = NULL,
  width = NULL,
  jitter_seed = NA,
  violin_scale = "count",
  legend = NULL,
  legend.position = NULL,
  legend.title = NULL,
  legend.reverse = FALSE,
  legend.nrow = NULL,
  legend.barheight = 6,
  legend.barwidth = 1.5,
  legend.nbin = 300,
  legend.italic = FALSE,
  sankey_node_width = 0.15,
  sankey_node_whitespace = 0.03,
  sankey_alpha = 0.5,
  sankey_remove_axes = NULL,
  zoom = FALSE,
  sep = " / ",
  print = FALSE,
  text_factor = getOption("plot2.text_factor", 1),
  font = getOption("plot2.font"),
  theme = getOption("plot2.theme", "theme_minimal2"),
  background = getOption("plot2.colour_background", "white"),
  markdown = TRUE,
  data = NULL,
  ...
)

Arguments

.data

Data to plot. A single character value that is the name of a data frame in the calling environment (such as plot2("df"), as used in tool calling) plots that data frame; only the calling environment is searched, not the search path.

x

Plotting 'direction' for the X-axis. This can be:

A single variable from .data, such as x = column1
A function to calculate over one or more variables from .data, such as x = format(column1, "%Y"), or x = ifelse(column1 == "A", "Group A", "Other")
Multiple variables from .data, such as x = c(column1, column2, column2), or using selection helpers such as x = where(is.character) or x = starts_with("var_") (only allowed and required for Sankey plots using type = "sankey")

y

Values to use for plotting along the Y-axis. This can be:

A single variable from .data, such as y = column1
Multiple variables from .data, such as y = c(column1, column2) or y = c(name1 = column1, "name 2" = column2), or using selection helpers such as y = where(is.double) or y = starts_with("var_") (multiple variables only allowed if category is not set)
A function to calculate over .data returning a single value, such as y = n() for the row count, or based on other variables such as y = n_distinct(person_id), y = max(column1), or y = median(column2) / column3
A function to calculate over .data returning multiple values, such as y = quantile(column1, c(0.25, 0.75)) or y = range(age) (multiple values only allowed if category is not set)

category, facet

Plotting 'direction' (category is called 'fill' and 'colour' in ggplot2). This can be:

A single variable from .data, such as category = column1
A function to calculate over one or more variables from .data, such as category = median(column2) / column3, or facet = ifelse(column1 == "A", "Group A", "Other")
Multiple variables from .data, such as facet = c(column1, column2) (use sep to control the separator character)
One or more variables from .data using selection helpers, such as category = where(is.double) or facet = starts_with("var_")

The category can also be a date or date/time (class Date or POSIXt).

type

Type of visualisation to use. This can be:

A ggplot2 geom name or their abbreviation such as "col" and "point". All geoms are supported (including geom_blank()).

Full function names can be used (e.g., "geom_histogram"), but they can also be abbreviated (e.g., "h", "hist"). The following geoms can be abbreviated by their first character: area ("a"), boxplot ("b"), column ("c"), histogram ("h"), jitter ("j"), line ("l"), point ("p"), ribbon ("r"), and violin ("v"). Beeswarm plots ("beeswarm", short: "bs") require the ggbeeswarm package.

Please note: in ggplot2, 'bars' and 'columns' are equal, while it is common to many people that 'bars' are oriented horizontally and 'columns' are oriented vertically since Microsoft Excel has been using these terms this way for many years. For this reason, type = "bar" will set type = "col" and horizontal = TRUE.
One of these additional types:
- "barpercent" (short: "bp"), which is effectively a shortcut to set type = "col" and horizontal = TRUE and x.max_items = 10 and x.sort = "freq-desc" and datalabels.format = "%n (%p)".
- "dumbbell" (short: "d" or "db"), which sets type = "point" and horizontal = TRUE, and adds a line between the points (using geom_segment()). The line colour cannot be changed. This plot type is only possible when the category has two distinct values.
- "linedot" or "line-point" (short: "ld" or "lp"), which sets type = "line" and adds two point geoms using add_point(); one with large white dots and one with smaller dots using the colours set in colour. This is essentially equal to base R plot(..., type = "b") but with closed shapes.
- "back-to-back" (short: "b2b", or "butterfly", "tornado" or "population") creates a back-to-back plot, sometimes called a Tornado diagram, a butterfly plot, or a Population Pyramid. It uses facet to distinguish the left and right plots. Therefore, facet must be set to (a column containing) two unique values.
- "upset" or "UpSet" (short: "u") creates an UpSet plot, which requires x to contain multiple variables from .data that contain 0/1 or FALSE/TRUE values. It is also possible to provide y, on which then summarise_function will be applied.
- "sankey" (short: "s") creates a Sankey plot using category for the flows and requires x to contain multiple variables from .data. At default, it also sets x.expand = c(0.05, 0.05) and y.limits = c(NA, NA) and y.expand = c(0.01, 0.01). The so-called 'nodes' (the 'blocks' with text) are considered the datalabels, so you can set the text size and colour of the nodes using datalabels.size, datalabels.colour, and datalabels.colour_fill. The transparency of the flows can be set using sankey_alpha, and the width of the nodes can be set using sankey_node_width. Sankey plots can also be flipped using horizontal = TRUE, and the label angles can be set with datalabels.angle.
- "spider" or "radar" (short: "sp") creates a Spider plot using our novel CoordSpider coordinate system. Facets are supported, always with the same scale in all facets.
Left blank. In this case, the type will be determined automatically: "boxplot" if there is no X-axis or if the length of unique values per X-axis item is at least 3, "point" if both the y and x axes are numeric, and the option "plot2.default_type" otherwise (which defaults to "col", and accepts any value of type, such as "line" or "l"). Use type = "blank" or type = "geom_blank" to not add a geom.

title, subtitle, caption, tag, x.title, y.title, category.title, legend.title, y_secondary.title

A title to use. This can be:

A character, which supports markdown by using md_to_expression() internally if markdown = TRUE (which is the default)
A function to calculate over .data, such as title = paste("Based on n =", n_distinct(person_id), "individuals") or subtitle = paste("Total rows:", n()), see Examples
An expression, e.g. using parse(text = "...")

The category.title defaults to TRUE if the legend items are numeric. category.title and legend.title are the same setting; use one of them.

title.linelength

Maximum number of characters per line in the title, before a linebreak occurs.

title.colour

Text colour of the title.

subtitle.linelength

Maximum number of characters per line in the subtitle, before a linebreak occurs.

subtitle.colour

Text colour of the subtitle.

na.replace

Character to put in place of NA values if na.rm = FALSE.

na.rm

Remove NA values from showing in the plot.

facet.position, facet.colour_fill, facet.bold, facet.italic, facet.size, facet.margin, facet.repeat_labels_x, facet.repeat_labels_y, facet.drop, facet.nrow, facet.relative

Additional settings for the plotting direction facet. facet.repeat_labels_x and facet.repeat_labels_y repeat the axis labels in every facet, which also makes the scales of that axis free (unless facet.fixed_x or facet.fixed_y is TRUE).

facet.fixed_y

A logical to indicate whether all y scales should have the same limits. If not set, this depends on the data: TRUE only if the coefficient of variation (standard deviation divided by mean) of the maximum y values per facet is less than 25%, otherwise every facet gets its own y scale. Both outcomes are reported in a message. Since this can change when the data change (e.g., next month's data), set facet.fixed_y = TRUE when panels must be compared.

facet.fixed_x

A logical to indicate whether all x scales should have the same breaks. This acts like the inverse of x.drop. If not set and x is numeric, the same rule as for facet.fixed_y applies to the maximum x values per facet.

x.date_breaks

Breaks to use when the X-axis contains dates, will be determined automatically if left blank. This accepts values such as "1 day" and "2 years".

x.date_labels

Labels to use when the X-axis contains dates, will be determined automatically if left blank. This accepts 'Excel' date-language such as "d mmmm yyyy": d/dd (day), ddd/dddd (weekday name), mm (month number), mmm/mmmm (month name), yy/yyyy (year), q/qq (quarter: 1 or Q1), w/ww (ISO week), h/hh (hour), MM (minutes), ss (seconds). Other words are kept as they are (e.g. "Week ww"), and text in single quotes is always kept literally (e.g. "'d' d"). A format containing % is used as a strftime() format. Month and weekday names follow the option plot2.language if set, and the system locale otherwise.

x.date_remove_years

A logical to indicate whether the years of all x values must be unified. This will set the years of all x values to 1970 if the data does not contain a leap year, and to 1972 otherwise. This allows to plot years on the category while maintaining a date range on x. The default is FALSE, unless category contains all years present in x.

category.focus

A value of category that should be highlighted, meaning that all other values in category will be greyed out. This can also be a numeric value between 1 and the length of unique values of category, e.g. category.focus = 2 to focus on the second legend item.

colour

Colour(s) to set, will be evaluated with get_colour() if set. This can also be one of the viridis colours with automatic implementation for any plot: "viridis", "magma", "inferno", "plasma", "cividis", "rocket", "mako" or "turbo". Also, this can be a named vector to match values of category, see Examples. Using a named vector can be used to manually sort the values of category.

colour_fill

Colour(s) to be used for filling, will be determined automatically if left blank and will be evaluated with get_colour().

colour_transparency

Amount of transparency for colour/colour_fill (0 = solid, 1 = fully transparent).

x.label_angle

Angle to use for the X-axis in a counter-clockwise direction (i.e., a value of 90 will orient the axis labels from bottom to top, a value of 270 will orient the axis labels from top to bottom).

x.label_align

Alignment for the X-axis between 0 (left aligned) and 1 (right aligned). If not set, rotated labels (see x.label_angle) are anchored at their ticks with guide_axis(angle = ), for every angle and also when horizontal = TRUE.

x.label_italic

logical to indicate whether the x labels should be in italics.

x.remove, y.remove

A logical to indicate whether the axis labels and title should be removed.

x.position, y.position

Position of the axis.

x.max_items, category.max_items, facet.max_items

Maximum number of items to show, including the item that holds the rest: x.max_items = 5 shows four values and the rest, see x.max_txt. Defaults to infinite. Left NULL (default for x.max_items), this is infinite, except for type = "barpercent" and back-to-back plots (10). The remaining values are grouped into one item and summarised using the summarise_function function. Please note: the sorting will be applied first, so that e.g. x.sort = "freq-desc" with x.max_items = 5 shows the four most frequent values of the X-axis and the rest.

x.max_txt, category.max_txt, facet.max_txt

The text of the item that holds the values beyond *.max_items. The placeholder %n is replaced with the number of values in that item, and %p with their share of the total of y (or of the rows, when y is not numeric). For example, the default "(rest, x%n)" gives "(rest, x4)" for four values.

x.breaks, y.breaks

A breaks function or numeric vector to use for the axis.

x.n_breaks, y.n_breaks

Number of breaks, only useful if x.breaks or y.breaks respectively is NULL.

x.transform, y.transform, category.transform

A transformation function to use, e.g. "log2". This can be: "asinh", "asn", "atanh", "boxcox", "compose", "date", "exp", "hms", "identity", "log", "log10", "log1p", "log2", "logit", "modulus", "probability", "probit", "pseudo_log", "reciprocal", "reverse", "sqrt", "time", "timespan", "yj".

x.expand, y.expand

expansion to use for the axis, can be length 1 or 2. Left NULL (default), they are determined automatically: x.expand adds half a category on both sides of a discrete axis and a proportion of the range on a continuous axis, y.expand adds 25% above the data; both are 0 for sf objects or when the limits are set. An explicit value is used as given.

x.limits, y.limits

Limits to use for the axis, must be length 2. Use NA for the highest or lowest value in the data, e.g. y.limits = c(0, NA) to have the y scale start at zero. For a discrete x axis, x.limits are the items to show, in this order.

x.labels, y.labels, y_secondary.labels

A labels function or character vector to use for the axis.

x.character

A logical to indicate whether the values of the X-axis should be forced to character. If not set, whole numbers that look like years (2000 to 2050) or months (exactly 1 to 12) become character when type is not set or is a discrete type such as "col" or "boxplot", with a message; for continuous types such as "point", "line" and "area" they stay numeric. Numeric values also become character for other discrete types. Use x.character = FALSE to keep numeric values.

x.drop

logical to indicate whether factor levels should be dropped.

x.zoom, y.zoom

A logical to indicate if the axis should be zoomed on the data, i.e., cover the range of the data instead of starting at 0. Left NULL (default for x.zoom), the x axis is zoomed for plot types with a continuous x axis (such as histograms) only. This sets x.limits = c(NA, NA) (or y.limits) and, unless x.expand (or y.expand) is set, a 5% expansion on both sides, like ggplot2 does. It has no effect if x.limits (or y.limits) is set.

y.24h

A logical to indicate whether the y labels and breaks should be formatted as 24-hour sequences.

y.age

A logical to indicate whether the y labels and breaks should be formatted as ages in years.

y.scientific, y_secondary.scientific

A logical to indicate whether the y labels should be formatted in scientific notation. If not set, scientific notation is used only for labels of one billion (1e9) or more, or for non-zero labels below 0.0001; other labels use big.mark, e.g. 2 000 000.

y.percent, y_secondary.percent

A logical to indicate whether the y labels should be formatted as percentages. Left NULL (default for y.percent), percentages are used for density plots only.

y.percent_break

A value on which the Y-axis should have breaks if y.percent = TRUE. Left NULL (default), this is 0.1 (10%), widened to keep about 10 labels and narrowed if that would give three or fewer labels. An explicit value is used as given.

y_secondary

Values to use for plotting along the secondary Y-axis. This functionality is poorly supported by ggplot2 and might give unexpected results. Setting the secondary Y-axis will set the colour to the axis titles.

y_secondary.type

Type of the secondary series, see type. The default (NULL) is a line with points if the primary type is a column or bar plot, so that the secondary series does not cover the bars, and the primary type otherwise.

y_secondary.colour, y_secondary.colour_fill

Colours to set for the secondary Y-axis, will be evaluated with get_colour(). Left NULL (default), the second colour of the plot2.colour option is used.

category.type

Type of the category, one or more of: "colour" (default), "shape", "size", "linetype", "linewidth", "alpha". There is no need to set "fill", as plot2 handles colour-setting internally and determines automatically whether the colour or fill aesthetic must be used.

category.labels, category.percent, category.breaks, category.expand, category.midpoint

Settings for the plotting direction category.

category.limits

Limits to use for a numeric category, must be length 2. Use NA for the highest or lowest value in the data, e.g. category.limits = c(0, NA) to have the scale start at zero. If not set, the limits are the range of the data, rounded outwards to pretty values with pretty() (e.g., values from 12 to 87 give limits 0 to 100). With three colours and category.midpoint set, the limits are made symmetric around the midpoint.

category.date_breaks

Breaks to use when the category contains dates, will be determined automatically if left blank. This will be passed on to seq.Date(by = ...) and thus can be: a number, taken to be in days, or a character string containing one of "day", "week", "month", "quarter" or "year" (optionally preceded by an integer and a space, and/or followed by "s").

category.date_labels

Labels to use when the category contains dates, will be determined automatically if left blank. This accepts 'Excel' date-language such as "d mmmm yyyy".

category.character

A logical to indicate whether the values of the category should be forced to character. If not set, whole numbers that look like years (2000 to 2050) or months (exactly 1 to 12) become character, with a message. Use category.character = FALSE to keep numeric values.

x.sort, category.sort, facet.sort

Sorting of the plotting direction. category.sort and facet.sort default to TRUE; x.sort defaults to NULL, which sorts text and factors as TRUE and leaves continuous values (such as dates and numbers) unsorted. Applying one of the sorting methods will transform the values to an ordered factor, which ggplot2 uses to orient the data. Valid values are:

A manual vector of values
TRUE: sort factors on their levels, otherwise sort ascending on alphabet, while maintaining numbers in the text (numeric sort)
FALSE: sort according to the order in the data
NULL: sort automatically, as described above
"asc", "alpha" or "alpha-asc": sort as TRUE
"desc" or "alpha-desc": sort factors on their reversed levels, otherwise sort descending on alphabet, while maintaining numbers in the text (numeric sort)
"order" or "inorder": sort as FALSE
"freq" or "freq-desc": sort descending according to the frequencies of y computed by summarise_function (highest value first)
"freq-asc": sort ascending according to the frequencies of y computed by summarise_function (lowest value first)

x.complete, category.complete, facet.complete

A value to complete the data. This makes use of tidyr::full_seq() and tidyr::complete(). For example, using x.complete = 0 will apply data |> complete(x = full_seq(x, ...), fill = list(y = 0)), crossed with the values of category and facet if set. Numeric values and dates are completed with tidyr::full_seq() at their observed resolution, date-times at their observed interval, and character values and factors with their unique values or levels. Using value TRUE (e.g., x.complete = TRUE) is identical to using value 0.

datalabels

Values to show as datalabels, see also datalabels.format. This can be:

NULL (default): automatic. This shows the values of y in column-type plots, or when plotting spatial 'sf' data, the values of the first column, with a maximum of 25 labels and none for continuous plot types such as lines and points.
TRUE to always show the values of y, without the maximum of 25 labels, or FALSE to remove datalabels
A function to calculate over .data, such as datalabels = paste(round(column1), "\n", column2)

datalabels.round

Number of digits to round the datalabels, applies to both "%n" and "%p" for replacement (see datalabels.format).

datalabels.format

Format to use for datalabels. This can be a function (such as euros()) or a text. For the text, "%n" will be replaced by the count number, and "%p" will be replaced by the percentage (see datalabels.percent_by). Use datalabels.format = NULL to not transform the datalabels.

datalabels.colour, datalabels.colour_fill, datalabels.size, datalabels.angle, datalabels.lineheight

Settings for the datalabels. Rotated datalabels (datalabels.angle) start or end at their bar with a small fixed gap; if y.expand and y.limits are not set, the value axis gets extra room for long rotated labels, estimated for a figure of about 6 by 4 inches.

datalabels.percent_by

Group within which the "%p" percentages of datalabels.format sum to 100%. This can be:

"x" (default): within each x value and facet, so that the stacked or dodged bars of one x value sum to 100%. Without a category, each x value has a single bar, so this is the same as "facet".
"category": within each category and facet
"facet": within each facet
"total": over all data in the plot

decimal.mark

Decimal mark, defaults to dec_mark().

big.mark

Thousands separator, defaults to big_mark().

summarise_function

A function to use if the data has to be summarised, see Examples. Left NULL (default), this is sum(), except when y contains several columns (then the values are not summarised). Use FALSE to not summarise the data at all.

stacked

A logical to indicate that values must be stacked, or "fill" to stack them to 100% (the same as stacked_fill = TRUE).

stacked_fill

A logical to indicate that values must be filled (i.e., stacked to 100%). If TRUE, this takes precedence over stacked.

horizontal

A logical to turn the plot 90 degrees using coord_flip(). This option also updates some theme options, so that e.g., x.label_italic will still apply to the original X-axis.

reverse

A logical to reverse the values of category. Use legend.reverse to reverse the legend of category.

smooth

A logical to add a smooth. In histograms, this will add the density count as an overlaying line (default: TRUE). In all other cases, a smooth will be added using geom_smooth() (default: FALSE).

smooth.method, smooth.formula, smooth.se, smooth.level, smooth.alpha, smooth.linewidth, smooth.linetype, smooth.colour

Settings for smooth.

size

Size of the geom. Defaults to 2 for geoms point, jitter, and beeswarm; 4 for UpSet plots (using type = "upset"); 5 for dumbbell plots (using type = "dumbbell"); 0.75 otherwise.

linetype

Linetype of the geom, only suitable for geoms that draw lines. Possible options are:

"solid" (default)
"dashed"
"dotted"
"dotdash"
"longdash"
"twodash"

Both numbers and text are supported, e.g., linetype = 2 and linetype = "dashed" yield the same results.

linewidth

Linewidth of the geom, only suitable for geoms that draw lines. Defaults to:

0.10 for sf
0.25 for geoms that are continuous and have fills (such as area), except for geoms boxplot and violin
1.0 for dumbbell plots (using type = "dumbbell") and UpSet plots (using type = "upset")
0.5 otherwise (such as geoms line, boxplot, violin, and histogram)

binwidth

Width of bins (only useful for type = "histogram"), can be specified as a numeric value or as a function that calculates width from x, see geom_histogram(). It defaults to approx. diff(range(x)) divided by 12 to 22 based on the data.

width

Width of the geom. Defaults to 0.66 for geoms boxplot, violin and jitter, and to 0.5 otherwise.

jitter_seed

Seed (randomisation factor) to be set when using type = "jitter".

violin_scale

Scale to be set when using type = "violin", can also be set to "area".

legend

A logical to show (TRUE) or hide (FALSE) the legend, i.e. the box explaining the values of category. The default NULL shows it when the plot has a category.

legend.position

Position of the legend. This can be:

"top", "right", "bottom" or "left", can be abbreviated
"inside", to place the legend inside the plot area at ggplot2's default position
Two numbers between 0 and 1, to place the legend inside the plot area at these relative coordinates, e.g. c(0.95, 0.95) for the top right
"none", FALSE or NA to remove the legend

The default NULL sets it automatically: "right" for numeric category values and 'sf' plots, and "top" otherwise.

legend.reverse, legend.barheight, legend.barwidth, legend.nbin, legend.italic, legend.nrow

Other settings for the legend.

sankey_node_width

Width of the vertical nodes in a Sankey plot.

sankey_node_whitespace

Whitespace between the nodes in a Sankey plot.

sankey_alpha

Alpha of the flows in a Sankey plot.

sankey_remove_axes

Logical to indicate whether all axes must be removed in a Sankey plot.

zoom

A logical to indicate if the plot should be scaled to the data, i.e., not having the x and y axes to start at 0. This is identical to setting x.zoom = TRUE and y.zoom = TRUE.

sep

Separator character to use if multiple columns are given to either of the three directions: x, category and facet, e.g. facet = c(column1, column2).

print

A logical to indicate if the result should be printed instead of just returned.

text_factor

Text factor to use, which will apply to all texts shown in the plot.

font

Font (family) to use, can be set with options(plot2.font = "..."). Can be any installed system font or any of the > 1400 font names from Google Fonts. Custom fonts require the showtext, showtextdb and sysfonts packages. A Google Font that is not installed is downloaded once into the user cache folder of plot2 (tools::R_user_dir("plot2", "cache")); fonts that have not been used for 90 days are removed from it automatically. The font is applied only while the plot is drawn, so other plots in the session are not affected. When using custom fonts in R Markdown, be sure to set the chunk option fig.showtext = TRUE, otherwise an informative error will be generated.

theme

A valid ggplot2 theme to apply, or NULL to use the default theme_grey(). This argument accepts themes (e.g., theme_bw()), functions (e.g., theme_bw) and characters themes (e.g., "theme_bw"). A partial theme, such as theme(panel.background = element_blank()), is added on top of theme_minimal2(), and its elements take precedence over plot2's own settings. The default is theme_minimal2(), but can be set with options(plot2.theme = "...").

background

The background colour of the entire plot, can also be NA to remove it. Will be evaluated with get_colour(). Only applies when theme is not NULL.

markdown

A logical to turn all labels and titles into plotmath expressions, by converting common markdown language using the md_to_expression() function (defaults to TRUE).

data

Data for formula notation only, e.g. plot2(mpg ~ hp, data = mtcars), which is the same as mtcars |> plot2(mpg ~ hp). A formula takes the form y ~ x + category + facet or y ~ x + category | facet; terms can be expressions, such as log(mpg) ~ hp. For other input, data gives an error: pass the data as the first argument instead.

...

Any argument to give to the geom. This will override automatically-set settings for the geom.

Value

A ggplot object.

Details

Plotting directions

A plot has up to four directions: x, y, category (the colours, called 'colour' and 'fill' in ggplot2) and facet (separate panels). Each can be a column or a calculation on the data. Groups of x, category and facet that contain several values are summarised with summarise_function, and each direction has its own settings, such as x.sort and category.sort for the order of the values, and x.max_items and category.max_items to show only the first values and combine the rest into one. Directions that are not given are chosen from the data, see Automatic choices below.

Plot types

type accepts every ggplot2 geom and adds the types "barpercent", "linedot", "dumbbell", "back-to-back", "sankey", "spider" and "upset". Without type, the type follows from the data. See vignette("supported_types", package = "plot2") for every type with its data requirements.

The result

The result is an ordinary ggplot object: it can be extended with + and any ggplot2 function, with the add_*() functions, which add layers with unquoted variables, and its layers can be reordered with move_layer(). The default theme is theme_minimal2(), and the colours are set with colour and colour_fill, which accept any value of get_colour().

Text and numbers

Titles and labels support markdown (such as "**bold**" and "*italic*") and calculations on the data, such as title = paste("Patients in", min(format(date, "%Y"))). The font can be any installed system font or a Google Font (font). Numbers use the decimal and big marks of dec_mark() and big_mark(), and the language of generated text follows the plot2.language option; see plot2-options for all options.

Further reading

vignette("plot2", package = "plot2") introduces the package step by step. The package website has articles on colours and themes, maps, fonts, and plot2() for ggplot2 users. create_interactively() builds a plot in a Shiny app and shows the code that creates it.

The generic plot2() currently has 153 arguments, all with a default value except .data and ....

Automatic choices

Like base R's plot(), plot2() decides what is not given. Every decision follows a fixed rule, is announced in a message (in interactive sessions, see the plot2.silent option in plot2-options), and is overridden by setting the argument concerned. The rules are:

What to plot

y: a column named n, otherwise the first numeric column (the second one if x is also chosen automatically); y = n() to count rows if the data have no numeric column, or for type = "barpercent" and Sankey plots.
x: the first numeric column if there is no x and the data have several numeric columns, otherwise the first column not used elsewhere.
category: the first character or factor column if neither x nor category is given and the data have more than two columns; for 'sf' data, the first numeric column.
type: a boxplot if there is no x, or if every group of x, category and facet has at least three values; points if x and y are both numeric; a histogram if x is the only numeric column; otherwise the plot2.default_type option (columns). A matrix with identical row and column names becomes a heatmap ("tile").
summarise_function: sum() if groups contain several values (see summarise_function).

Axes and scales

x.character and category.character: TRUE if the values are whole years from 2000 to 2050, or exactly the months 1 to 12 (for x only in discrete plot types); x.character also if x is numeric in a discrete plot type, or if x.sort is set.
x.date_remove_years: TRUE if category contains the years of a date x.
x.date_breaks, x.date_labels, category.date_breaks and category.date_labels: based on the range of the dates.
facet.fixed_x and facet.fixed_y: the same scale in all facets if the maxima per facet differ by less than 25% (coefficient of variation), free scales otherwise; facet.repeat_labels_y = FALSE with a fixed y scale.
x.expand and y.expand: 0 if the limits are set.
y.percent: TRUE for density plots; y.percent_break: about 10 labels.
y.scientific: TRUE if a non-zero label would be 1e9 or larger, or smaller than 1e-4.
category.midpoint: the centre of the category scale, for three colours.
binwidth: based on the data for histograms, which also get smooth = TRUE.

Labels and layout

datalabels: the values of y in column-type plots, at most 25 and none for continuous plot types such as points and lines; for 'sf' data, the first character column, centred on the areas (datalabels.centroid = TRUE).
legend.position: "right" for numeric categories and 'sf' data, "top" otherwise; numeric categories also get a legend title.
sankey_remove_axes: TRUE for Sankey plots.

Argument names

The arguments follow one naming rule: <argument>.<setting>, where the part before the dot is itself an argument of plot2() (x, y, category, facet, y_secondary, title, subtitle, datalabels, smooth or legend) and the setting is written in snake case, e.g. x.label_angle or facet.repeat_labels_x. All other arguments are plain snake case, e.g. colour_fill, sankey_node_width or jitter_seed. The exceptions are names known from base R: na.rm, na.replace, decimal.mark and big.mark.

A default of NULL means that plot2() determines the value automatically, often announced in a message. An explicit value is always used as given.

Examples

# without variables, plot2() determines them: two numeric columns give points
iris |>
  plot2()
#> Using x = Sepal.Length
#> Using y = Sepal.Width
#> Using category = Species
#> Using type = "point" since both axes are numeric


# the arguments are in this order: x, y, category, facet
iris |>
  plot2(Sepal.Length, Sepal.Width, Petal.Length, Species,
        colour = "viridis")
#> Using type = "point" since both axes are numeric
#> Assuming facet.fixed_x = TRUE since the three x scales differ by less than 25%
#> Assuming facet.fixed_y = TRUE since the three y scales differ by less than 25%
#> Assuming facet.repeat_labels_y = FALSE since y has fixed scales


# y can be several columns, also with selection helpers
iris |>
  plot2(x = Species, y = where(is.double))
#> Using y = c(Petal.Length, Petal.Width, Sepal.Length, Sepal.Width)
#> Using type = "boxplot" since all groups in Species and category contain at least three values


# y can be any calculation; datalabels are added for columns
admitted_patients |>
  plot2(hospital, n(), gender)


# sorting and stacking are single arguments
admitted_patients |>
  plot2(hospital, n(), gender,
        x.sort = "freq-asc",
        stacked = TRUE)
#> Applying x.sort = "freq-asc" using summarise_function = sum


# set the plot type explicitly
admitted_patients |>
  plot2(hospital, median(age), gender, type = "dumbbell")


# titles support calculations too
admitted_patients |>
  plot2(age, type = "hist",
        title = paste("Based on n =", n_distinct(patient_id), "patients"))
#> Using type = "histogram"
#> Using binwidth = 6.4 based on data
#> Assuming smooth = TRUE for type = "histogram"


# a correlation matrix becomes a heatmap
mtcars |>
  cor() |>
  plot2()
#> Assuming type = "tile" since the matrix contains identical row and column names
#> Omitting printing of 121 datalabels - use datalabels = TRUE to force printing


# many more examples, including Sankey, spider and geographic plots,
# are in the vignettes:
# vignette("plot2", package = "plot2")
# vignette("supported_types", package = "plot2")

Create a `ggplot` in a Single Call