The sina plot is a data visualization chart suitable for plotting any single variable in a multiclass dataset. It is an enhanced jitter strip chart, where the width of the jitter is controlled by the density distribution of the data within each class.

stat_sina(mapping = NULL, data = NULL, geom = "sina",
  position = "dodge", scale = "area", method = "density",
  bw = "nrd0", kernel = "gaussian", maxwidth = NULL, adjust = 1,
  bin_limit = 1, binwidth = NULL, bins = NULL, seed = NA, ...,
  na.rm = FALSE, show.legend = NA, inherit.aes = TRUE)

geom_sina(mapping = NULL, data = NULL, stat = "sina",
  position = "dodge", ..., na.rm = FALSE, show.legend = NA,
  inherit.aes = TRUE)



Set of aesthetic mappings created by aes() or aes_(). If specified and inherit.aes = TRUE (the default), it is combined with the default mapping at the top level of the plot. You must supply mapping if there is no plot mapping.


The data to be displayed in this layer. There are three options:

If NULL, the default, the data is inherited from the plot data as specified in the call to ggplot().

A data.frame, or other object, will override the plot data. All objects will be fortified to produce a data frame. See fortify() for which variables will be created.

A function will be called with a single argument, the plot data. The return value must be a data.frame, and will be used as the layer data.


The geometric object to use display the data


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


How should each sina be scaled. Corresponds to the scale parameter in ggplot2::geom_violin()? Available are:

  • 'area' for scaling by the largest density/bin among the different sinas

  • 'count' as above, but in addition scales by the maximum number of points in the different sinas.

  • 'width' Only scale according to the maxwidth parameter

For backwards compatibility it can also be a logical with TRUE meaning area and FALSE meaning width


Choose the method to spread the samples within the same bin along the x-axis. Available methods: "density", "counts" (can be abbreviated, e.g. "d"). See Details.


the smoothing bandwidth to be used. The kernels are scaled such that this is the standard deviation of the smoothing kernel. (Note this differs from the reference books cited below, and from S-PLUS.)

bw can also be a character string giving a rule to choose the bandwidth. See bw.nrd.
The default, "nrd0", has remained the default for historical and compatibility reasons, rather than as a general recommendation, where e.g., "SJ" would rather fit, see also Venables and Ripley (2002).

The specified (or computed) value of bw is multiplied by adjust.


a character string giving the smoothing kernel to be used. This must partially match one of "gaussian", "rectangular", "triangular", "epanechnikov", "biweight", "cosine" or "optcosine", with default "gaussian", and may be abbreviated to a unique prefix (single letter).

"cosine" is smoother than "optcosine", which is the usual ‘cosine’ kernel in the literature and almost MSE-efficient. However, "cosine" is the version used by S.


Control the maximum width the points can spread into. Values between 0 and 1.


the bandwidth used is actually adjust*bw. This makes it easy to specify values like ‘half the default’ bandwidth.


If the samples within the same y-axis bin are more than bin_limit, the samples's X coordinates will be adjusted.


The width of the bins. The default is to use bins bins that cover the range of the data. You should always override this value, exploring multiple widths to find the best to illustrate the stories in your data.


Number of bins. Overridden by binwidth. Defaults to 50.


A seed to set for the jitter to ensure a reproducible plot


Other arguments passed on to layer(). These are often aesthetics, used to set an aesthetic to a fixed value, like colour = "red" or size = 3. They may also be parameters to the paired geom/stat.


If FALSE, the default, missing values are removed with a warning. If TRUE, missing values are silently removed.


logical. Should this layer be included in the legends? NA, the default, includes if any aesthetics are mapped. FALSE never includes, and TRUE always includes. It can also be a named logical vector to finely select the aesthetics to display.


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


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


There are two available ways to define the x-axis borders for the samples to spread within:

  • method == "density" A density kernel is estimated along the y-axis for every sample group, and the samples are spread within that curve. In effect this means that points will be positioned randomly within a violin plot with the same parameters.

  • method == "counts": The borders are defined by the number of samples that occupy the same bin.


geom_sina understand the following aesthetics (required aesthetics are in bold):

  • x

  • y

  • color

  • group

  • size

  • alpha

Computed variables


The density or sample counts per bin for each point


density scaled by the maximum density in each group


The number of points in the group the point belong to


ggplot(midwest, aes(state, area)) + geom_point()
# Boxplot and Violin plots convey information on the distribution but not the # number of samples, while Jitter does the opposite. ggplot(midwest, aes(state, area)) + geom_violin()
ggplot(midwest, aes(state, area)) + geom_jitter()
# Sina does both! ggplot(midwest, aes(state, area)) + geom_violin() + geom_sina()
p <- ggplot(midwest, aes(state, popdensity)) + scale_y_log10() p + geom_sina()
# Colour the points based on the data set's columns p + geom_sina(aes(colour = inmetro))
# Or any other way cols <- midwest$popdensity > 10000 p + geom_sina(colour = cols + 1L)
# Sina plots with continuous x: ggplot(midwest, aes(cut_width(area, 0.02), popdensity)) + geom_sina() + scale_y_log10()
### Sample gaussian distributions # Unimodal a <- rnorm(500, 6, 1) b <- rnorm(400, 5, 1.5) # Bimodal c <- c(rnorm(200, 3, .7), rnorm(50, 7, 0.4)) # Trimodal d <- c(rnorm(200, 2, 0.7), rnorm(300, 5.5, 0.4), rnorm(100, 8, 0.4)) df <- data.frame( 'Distribution' = c( rep('Unimodal 1', length(a)), rep('Unimodal 2', length(b)), rep('Bimodal', length(c)), rep('Trimodal', length(d)) ), 'Value' = c(a, b, c, d) ) # Reorder levels df$Distribution <- factor( df$Distribution, levels(df$Distribution)[c(3, 4, 1, 2)] ) p <- ggplot(df, aes(Distribution, Value)) p + geom_boxplot()
p + geom_violin() + geom_sina()
# By default, Sina plot scales the width of the class according to the width # of the class with the highest density. Turn group-wise scaling off with: p + geom_violin() + geom_sina(scale = FALSE)