x , y
|
the x and y arguments provide the x and y coordinates for the plot. Any reasonable way of defining the coordinates is acceptable; most likely the names of existing vectors or columns of data frames. See the ‘Examples’ section below, or the function xy.coords for details. If supplied separately, x and y must be of the same length.
|
…
|
other graphical parameters (see par ).
|
by
|
grouping variable(s). The default behaviour is for groups to be represented in the form of distinct colours, which will also trigger an automatic legend. (See legend below for customization options.) However, groups can also be presented through other plot parameters (e.g., pch or lty ) by passing an appropriate "by" keyword; see Examples. Note that continuous (i.e., gradient) colour legends are also supported if the user passes a numeric or integer to by . To group by multiple variables, wrap them with interaction .
|
facet
|
the faceting variable(s) that you want arrange separate plot windows by. Can be specified in various ways:
-
In "atomic" form, e.g. facet = fvar . To facet by multiple variables in atomic form, simply interact them, e.g. interaction(fvar1, fvar2) or factor(fvar1):factor(fvar2) .
-
As a one-sided formula, e.g. facet = ~fvar . Multiple variables can be specified in the formula RHS, e.g. ~fvar1 + fvar2 or ~fvar1:fvar2 . Note that these multi-variable cases are all treated equivalently and converted to interaction(fvar1, fvar2, …) internally. (No distinction is made between different types of binary operators, for example, and so f1+f2 is treated the same as f1:f2 , is treated the same as f1*f2 , etc.)
-
As a two-side formula, e.g. facet = fvar1 ~ fvar2 . In this case, the facet windows are arranged in a fixed grid layout, with the formula LHS defining the facet rows and the RHS defining the facet columns. At present only single variables on each side of the formula are well supported. (We don’t recommend trying to use multiple variables on either the LHS or RHS of the two-sided formula case.)
-
As a special “by” convenience keyword, in which case facets will match the grouping variable(s) passed to by above.
|
facet.args
|
an optional list of arguments for controlling faceting behaviour. (Ignored if facet is NULL.) Supported arguments are as follows:
-
nrow , ncol for overriding the default "square" facet window arrangement. Only one of these should be specified, but nrow will take precedence if both are specified together. Ignored if a two-sided formula is passed to the main facet argument, since the layout is arranged in a fixed grid.
-
fmar a vector of form c(b,l,t,r) for controlling the base margin between facets in terms of lines. Defaults to the value of tpar(“fmar”) , which should be c(1,1,1,1) , i.e. a single line of padding around each individual facet, assuming it hasn’t been overridden by the user as part their global tpar settings. Note some automatic adjustments are made for certain layouts, and depending on whether the plot is framed or not, to reduce excess whitespace. See tpar for more details.
-
cex , font , col , bg , border for adjusting the facet title text and background. Default values for these arguments are inherited from tpar (where they take a "facet." prefix, e.g. tpar(“facet.cex”) ). The latter function can also be used to set these features globally for all tinyplot plots.
|
data
|
a data.frame (or list) from which the variables in formula should be taken. A matrix is converted to a data frame.
|
type
|
character string or call to a type_*() function giving the type of plot desired.
-
NULL (default): Choose a sensible type for the type of x and y inputs (i.e., usually “p” ).
-
1-character values supported by plot :
-
“p” Points
-
“l” Lines
-
“b” Both points and lines
-
“c” Empty points joined by lines
-
“o” Overplotted points and lines
-
“s” Stair steps
-
“S” Stair steps
-
“h” Histogram-like vertical lines
-
“n” Empty plot over the extent of the data
-
tinyplot types:
-
“rect” , “segments” , or “polygon” : Equivalent to base R
-
“density” : Kernel density plot
-
“jitter” or type_jitter() : Jittered points
-
“polypath” or type_polypath()
-
“boxplot” or type_boxplot()
-
“histogram” or type_histogram()
-
“pointrange” or “errorbar” : segment intervals
-
“ribbon” or “area” for polygon intervals (where area plots are a special case of ribbon plots with ymin set to 0 and ymax set to y ; see below).
-
“lm” or type_lm() : Linear model fit
-
“glm” or type_glm() : Generalized linear model fit
-
“loess” or type_loess() : Local regression fit
|
xlim
|
the x limits (x1, x2) of the plot. Note that x1 > x2 is allowed and leads to a ‘reversed axis’. The default value, NULL, indicates that the range of the finite values to be plotted should be used.
|
ylim
|
the y limits of the plot.
|
log
|
a character string which contains "x" if the x axis is to be logarithmic, "y" if the y axis is to be logarithmic and "xy" or "yx" if both axes are to be logarithmic.
|
main
|
a main title for the plot, see also title .
|
sub
|
a subtitle for the plot.
|
xlab
|
a label for the x axis, defaults to a description of x.
|
ylab
|
a label for the y axis, defaults to a description of y.
|
ann
|
a logical value indicating whether the default annotation (title and x and y axis labels) should appear on the plot.
|
axes
|
logical or character. Should axes be drawn (TRUE or FALSE )? Or alternatively what type of axes should be drawn: “standard” (with axis, ticks, and labels; equivalent to TRUE ), “none” (no axes; equivalent to FALSE ), “ticks” (only ticks and labels without axis line), “labels” (only labels without ticks and axis line), “axis” (only axis line and labels but no ticks). To control this separately for the two axes, use the character specifications for xaxt and/or yaxt .
|
frame.plot
|
a logical indicating whether a box should be drawn around the plot. Can also use frame as an acceptable argument alias. The default is to draw a frame if both axis types (set via axes , xaxt , or yaxt ) include axis lines.
|
asp
|
the y/xy/x aspect ratio, see plot.window .
|
grid
|
argument for plotting a background panel grid, one of either:
-
a logical (i.e., TRUE to draw the grid), or
-
a panel grid plotting function like grid() . Note that this argument replaces the panel.first and panel.last arguments from base plot() and tries to make the process more seamless with better default behaviour. The default behaviour is determined by (and can be set globally through) the value of tpar(“grid”) .
|
palette
|
one of the following options:
-
NULL (default), in which case the palette will be chosen according to the class and cardinality of the "by" grouping variable. For non-ordered factors or strings with a reasonable number of groups, this will inherit directly from the user’s default palette (e.g., "R4"). In other cases, including ordered factors and high cardinality, the "Viridis" palette will be used instead. Note that a slightly restricted version of the "Viridis" palette—where extreme color values have been trimmed to improve visual perception—will be used for ordered factors and continuous variables. In the latter case of a continuous grouping variable, we also generate a gradient legend swatch.
-
A convenience string corresponding to one of the many palettes listed by either palette.pals() or hcl.pals() . Note that the string can be case-insensitive (e.g., "Okabe-Ito" and "okabe-ito" are both valid).
-
A palette-generating function. This can be "bare" (e.g., palette.colors ) or "closed" with a set of named arguments (e.g., palette.colors(palette = “Okabe-Ito”, alpha = 0.5) ). Note that any unnamed arguments will be ignored and the key n argument, denoting the number of colours, will automatically be spliced in as the number of groups.
|
legend
|
one of the following options:
-
NULL (default), in which case the legend will be determined by the grouping variable. If there is no group variable (i.e., by is NULL) then no legend is drawn. If a grouping variable is detected, then an automatic legend is drawn to the outer right of the plotting area. Note that the legend title and categories will automatically be inferred from the by argument and underlying data.
-
A convenience string indicating the legend position. The string should correspond to one of the position keywords supported by the base legend function, e.g. "right", "topleft", "bottom", etc. In addition, tinyplot supports adding a trailing exclamation point to these keywords, e.g. "right!", "topleft!", or "bottom!". This will place the legend outside the plotting area and adjust the margins of the plot accordingly. Finally, users can also turn off any legend printing by specifying "none".
-
Logical value, where TRUE corresponds to the default case above (same effect as specifying NULL) and FALSE turns the legend off (same effect as specifying "none").
-
A list or, equivalently, a dedicated legend() function with supported legend arguments, e.g. "bty", "horiz", and so forth.
|
pch
|
plotting "character", i.e., symbol to use. Character, integer, or vector of length equal to the number of categories in the by variable. See pch . In addition, users can supply a special pch = “by” convenience argument, in which case the characters will automatically loop over the number groups. This automatic looping will begin at the global character value (i.e., par(“pch”) ) and recycle as necessary.
|
lty
|
line type. Character, integer, or vector of length equal to the number of categories in the by variable. See lty . In addition, users can supply a special lty = “by” convenience argument, in which case the line type will automatically loop over the number groups. This automatic looping will begin at the global line type value (i.e., par(“lty”) ) and recycle as necessary.
|
lwd
|
line width. Numeric scalar or vector of length equal to the number of categories in the by variable. See lwd . In addition, users can supply a special lwd = “by” convenience argument, in which case the line width will automatically loop over the number of groups. This automatic looping will be centered at the global line width value (i.e.,
|
col
|
plotting color. Character, integer, or vector of length equal to the number of categories in the by variable. See col . Note that the default behaviour in tinyplot is to vary group colors along any variables declared in the by argument. Thus, specifying colors manually should not be necessary unless users wish to override the automatic colors produced by this grouping process. Typically, this would only be done if grouping features are deferred to some other graphical parameter (i.e., passing the "by" keyword to one of pch , lty , lwd , or bg ; see below.)
|
bg
|
background fill color for the open plot symbols 21:25 (see points.default ), as well as ribbon and area plot types. For the latter group—including filled density plots—an automatic alpha transparency adjustment will be applied (see the ribbon.alpha argument further below). Users can also supply either one of two special convenience arguments that will cause the background fill to inherit the automatic grouped coloring behaviour of col :
-
bg = “by” will insert a background fill that inherits the main color mappings from col .
-
by = <numeric[0,1]> (i.e., a numeric in the range [0,1] ) will insert a background fill that inherits the main color mapping(s) from col , but with added alpha-transparency.
For both of these convenience arguments, note that the (grouped) bg mappings will persist even if the (grouped) col defaults are themselves overridden. This can be useful if you want to preserve the grouped palette mappings by background fill but not boundary color, e.g. filled points. See examples.
|
fill
|
alias for bg . If non-NULL values for both bg and fill are provided, then the latter will be ignored in favour of the former.
|
alpha
|
a numeric in the range [0,1] for adjusting the alpha channel of the color palette, where 0 means transparent and 1 means opaque. Use fractional values, e.g. 0.5 for semi-transparency.
|
cex
|
character expansion. A numerical vector (can be a single value) giving the amount by which plotting characters and symbols should be scaled relative to the default. Note that NULL is equivalent to 1.0, while NA renders the characters invisible.
|
restore.par
|
a logical value indicating whether the par settings prior to calling tinyplot should be restored on exit. Defaults to FALSE, which makes it possible to add elements to the plot after it has been drawn. However, note the the outer margins of the graphics device may have been altered to make space for the tinyplot legend. Users can opt out of this persistent behaviour by setting to TRUE instead. See also get_saved_par for another option to recover the original par settings, as well as longer discussion about the trade-offs involved.
|
xmin , xmax , ymin , ymax
|
minimum and maximum coordinates of relevant area or interval plot types. Only used when the type argument is one of “rect” or “segments” (where all four min-max coordinates are required), or “pointrange” , “errorbar” , or “ribbon” (where only ymin and ymax required alongside x ).
|
ribbon.alpha
|
numeric factor modifying the opacity alpha of any ribbon shading; typically in [0, 1] . Only used when type = “ribbon” , or when the bg fill argument is specified in a density plot (since filled density plots are converted to ribbon plots internally). If an an applicable plot type is called but no explicit value is provided, then will default to tpar(“ribbon.alpha”) (i.e., probably 0.2 unless this has been overridden by the user in their global settings.)
|
add
|
logical. If TRUE, then elements are added to the current plot rather than drawing a new plot window. Note that the automatic legend for the added elements will be turned off.
|
file
|
character string giving the file path for writing a plot to disk. If specified, the plot will not be displayed interactively, but rather sent to the appropriate external graphics device (i.e., png , jpeg , pdf , or svg ). As a point of convenience, note that any global parameters held in (t)par are automatically carried over to the external device and don’t need to be reset (in contrast to the conventional base R approach that requires manually opening and closing the device). The device type is determined by the file extension at the end of the provided path, and must be one of ".png", ".jpg" (".jpeg"), ".pdf", or ".svg". (Other file types may be supported in the future.) The file dimensions can be controlled by the corresponding width and height arguments below, otherwise will fall back to the “file.width” and “file.height” values held in tpar (i.e., both defaulting to 7 inches, and where the default resolution for bitmap files is also specified as 300 DPI).
|
width
|
numeric giving the plot width in inches. Together with height , typically used in conjunction with the file argument above, overriding the default values held in tpar(“file.width”, “file.height”) . If either width or height is specified, but a corresponding file argument is not provided as well, then a new interactive graphics device dimensions will be opened along the given dimensions. Note that this interactive resizing may not work consistently from within an IDE like RStudio that has an integrated graphics windows.
|
height
|
numeric giving the plot height in inches. Same considerations as width (above) apply, e.g. will default to tpar(“file.height”) if not specified.
|
empty
|
logical indicating whether the interior plot region should be left empty. The default is FALSE . Setting to TRUE has a similar effect to invoking type = “n” above, except that any legend artifacts owing to a particular plot type (e.g., lines for type = “l” or squares for type = “area” ) will still be drawn correctly alongside the empty plot. In contrast,type = “n” implicitly assumes a scatterplot and so any legend will only depict points.
|
xaxt , yaxt
|
character specifying the type of x-axis and y-axis, respectively. See axes for the possible values.
|
flip
|
logical. Should the plot orientation be flipped, so that the y-axis is on the horizontal plane and the x-axis is on the vertical plane? Default is FALSE.
|
formula
|
a formula that optionally includes grouping variable(s) after a vertical bar, e.g. y ~ x | z . One-sided formulae are also permitted, e.g. ~ y | z . Multiple grouping variables can be specified in different ways, e.g. y ~ x | z1:z2 or y ~ x | z1 + z2 . (These two representations are treated as equivalent; both are parsed as interaction(z1, z2) internally.) Note that the formula and x arguments should not be specified in the same call.
|
subset , na.action , drop.unused.levels
|
arguments passed to model.frame when extracting the data from formula and data .
|