bin: Bin in 1d
In hadley/ggstat: Statistical Computations for Visualisation

Description Usage Arguments Floating point Examples

Bin a numeric vector and count how many observations fall in each bin. Supports weights so that you can re-bin pre-binned data.

bin_fixed(x, width = NULL, center = NULL, boundary = NULL,
  origin = NULL, terminus = NULL, bins = 30, pad = FALSE,
  closed = c("right", "left"))

bin_breaks(breaks, closed = c("right", "left"))

bin_date(x, bins = 30, closed = c("right", "left"))

`x`	A numeric vector to guess parameters from.
`width`	(Positive real). The width of a bin. For S3 objects, the interpretation of width depends on the interpretation of the underlying numeric vector. For example, for dates, 1 = 1 day; for times 1 = 1 second; and for difftime, the units vary. If `NULL`, the `width` will be derived from the data, picking approximately `bins` bins with nice widths. You should always override this value, exploring multiple widths to find the best to illustrate the stories in your data.
`boundary, center`	Set the position of the first bin by specifying the position of either a boundary or the center of a bin. For example, you can always center the bins on integers with `center = 0` regardless of where the first bin actually falls. Think of binning as tiling the real line into a infinite sequence of intervals. `center` and `boundary` set the position of one of those intervals.
`origin, terminus`	The locations of the left-most and right-most bins. Any values outside this range will be treated as missing. You should usually leave `origin` as `NULL` so that it is automatically computed from `center` and `boundary`.
`bins`	Number of bins to use if not specified. Pretty bin sizes are preferred over matching this value exactly.
`pad`	If `TRUE`, adds empty bins at either end of x. This ensures frequency polygons touch 0 outside the range of x. Defaults to `FALSE`.
`closed`	One of `"right"` or `"left"` indicating whether the bin interval is left-closed (i.e. [a, b)), or right-closed (i.e. (a, b]).
`breaks`	A numeric vector of break points.

If a point is less than binwidth / 10^8 from the boundary between two bins, it is shifted to fall in the bin with the closest "closed" side.

x <- runif(1e6)
compute_stat(bin_fixed(x), x)
compute_stat(bin_fixed(x, width = 0.25), x)
compute_stat(bin_breaks(c(0, 0.1, 0.9, 1)), x)

# Can also create fixed bins without data, if you supply the origin
# terminus, and width
bin_fixed(origin = 0, terminus = 1, width = 0.25)

bin_fixed(x, bins = 37)

# Bin other types of object
x1 <- Sys.time() + runif(1000) * 60
compute_stat(bin_date(x1), x1)
x2 <- Sys.Date() + sample(30, 10)
compute_stat(bin_date(x2), x2)

# For fixed bin width, performance scales linearly with the size of x.
x <- runif(1e7)
system.time(compute_stat(bin_fixed(x, width = 1e-1), x))
system.time(compute_stat(bin_fixed(x, width = 1e-2), x))
system.time(compute_stat(bin_fixed(x, width = 1e-5), x))

# For arbitrary breaks, performance scales linearly with x and
# logarthmically with the number of bins.
system.time(compute_stat(bin_breaks(seq(0, 1, length = 10)), x))
system.time(compute_stat(bin_breaks(seq(0, 1, length = 100)), x))
system.time(compute_stat(bin_breaks(seq(0, 1, length = 1000)), x))