Use optimization routines to find good locations for point labels without overlaps.
pointLabel(x, y = NULL, labels = seq(along = x), cex = 1,
method = c("SANN", "GA"),
allowSmallOverlap = FALSE,
trace = FALSE,
doPlot = TRUE,
...)
as with plot.default
, these provide the x and y coordinates for
the point labels. Any reasonable way of defining the coordinates is
acceptable. See the function xy.coords
for details.
as with text
, a character vector or expression specifying the text to be
written. An attempt is made to coerce other language objects
(names and calls) to expressions, and vectors and other
classed objects to character vectors by as.character
.
numeric character expansion factor as with text
.
the optimization method, either “SANN” for simulated annealing (the default) or “GA” for a genetic algorithm.
logical; if TRUE
, labels are allowed
a small overlap. The overlap allowed is 2% of the diagonal
distance of the plot area.
logical; if TRUE
, status updates are given as the optimization algorithms
progress.
logical; if TRUE
, the labels are plotted on the
existing graph with text
.
arguments passed along to text
to specify
labeling parameters such as col
.
An xy
list giving the x
and y
positions of the
label as would be placed by text(xy, labels)
.
Eight positions are candidates for label placement, either horizontally, vertically, or diagonally offset from the points. The default position for labels is the top right diagonal relative to the point (considered the preferred label position).
With the default settings, simulating annealing solves faster than the genetic algorithm. It is an open question as to which settles into a global optimum the best (both algorithms have parameters that may be tweaked).
The label positioning problem is NP-hard (nondeterministic polynomial-time hard). Placement becomes difficult and slows considerably with large numbers of points. This function places all labels, whether overlaps occur or not. Some placement algorithms remove labels that overlap.
Note that only cex
is used to calculate string width and
height (using strwidth
and strheight
), so passing a
different font may corrupt the label dimensions. You could get around
this by adjusting the font parameters with par
prior to running
this function.
https://en.wikipedia.org/wiki/Automatic_label_placement
http://i11www.iti.uni-karlsruhe.de/map-labeling/bibliography/
http://www.eecs.harvard.edu/~shieber/Projects/Carto/carto.html
http://www.szoraster.com/Cartography/PracticalExperience.htm
The genetic algorithm code was adapted from the python code at
https://meta.wikimedia.org/wiki/Map_generator.
The simulated annealing code follows the algorithm and guidelines in:
Jon Christensen, Joe Marks, and Stuart Shieber. Placing text labels on maps and diagrams. In Paul Heckbert, editor, Graphics Gems IV, pages 497-504. Academic Press, Boston, MA, 1994. http://www.eecs.harvard.edu/~shieber/Biblio/Papers/jc.label.pdf
text
, thigmophobe.labels
in package plotrix
# NOT RUN {
n <- 50
x <- rnorm(n)*10
y <- rnorm(n)*10
plot(x, y, col = "red", pch = 20)
pointLabel(x, y, as.character(round(x,5)), offset = 0, cex = .7)
plot(x, y, col = "red", pch = 20)
pointLabel(x, y, expression(over(alpha, beta[123])), offset = 0, cex = .8)
# }
Run the code above in your browser using DataLab