| colPoints {berryFunctions} | R Documentation |
Points colored relative to third dimension
Description
Draw colored points for 3D-data in a 2D-plane. Color is relative to third
dimension, by different classification methods. Can take 3 vectors or, as in
image, 2 vectors and a matrix for z.
Adding points after smallPlot is called for the legend may be
incorrect if the original function messes with the graph margins,
see the note in colPointsLegend.
Usage
colPoints(
x,
y,
z,
data,
add = TRUE,
col = seqPal(100),
col2 = c(NA, "grey", "black"),
Range = range(z, finite = TRUE),
method = "linear",
breaks = length(col),
sdlab = 1,
legend = TRUE,
legargs = NULL,
lines = FALSE,
nint = 30,
xlab = gsub("\"", "", deparse(substitute(x))),
ylab = gsub("\"", "", deparse(substitute(y))),
zlab = gsub("\"", "", deparse(substitute(z))),
axes = TRUE,
log = "",
las = 1,
bglines = NULL,
pch = 16,
x1 = 0.6,
y1 = ifelse(horizontal, 0.88, 0.3),
x2 = 0.99,
y2 = 0.99,
density = NULL,
horizontal = TRUE,
quiet = FALSE,
...
)
Arguments
x, y |
Vectors with coordinates of the points to be drawn |
z |
z values belonging to coordinates. Vector or matrix with the color-defining height values |
data |
Optional: data.frame with the column names as given by x,y and z. |
add |
Logical. Should the points be added to current (existing!) plot? If FALSE, a new plot is started. DEFAULT: TRUE (It's called colPoints, after all) |
col |
Vector of colors to be used. DEFAULT: 100 colors from sequential
palette |
col2 |
Color for points where z is NA, or lower / higher than |
Range |
Ends of color bar. If NULL, it is again the DEFAULT: range(z, finite=TRUE) |
method |
Classification method (partial matching is performed),
see |
breaks |
Specification for method, see |
sdlab |
Type of label and breakpoints if |
legend |
Logical. Should a |
legargs |
List. Arguments passed to |
lines |
Logical. Should lines be drawn instead of / underneath the points?
(color of each |
nint |
Numeric of length 1. Number of interpolation points between each
coordinate if |
xlab, ylab, zlab |
X axis label, y axis label, |
axes, las |
Draw axes? Label Axis Style. Only used when add=FALSE.
See |
log |
Logarithmic axes with log="y", "xy" or "x". For logarithmic colorscale, see method="log". DEFAULT: "" |
bglines |
If not NULL, passed to |
pch |
Point CHaracter. See |
x1, x2, y1, y2 |
Relative coordinates [0:1] of inset plot, see |
density |
Arguments for density line in |
horizontal |
Logical passed to |
quiet |
Turn off warnings? DEFAULT: FALSE |
... |
Further graphical arguments passed to |
Value
Invisible list of values that can be passed to colPointsLegend or colPointsHist.
Note
Rstudio scales graphics really badly, so don't expect the right legend width out of the box if you use Rstudio!
Exporting via png("myplot.png", 600,400); colPoints(x,y,z); dev.off() usually works much better
Author(s)
Berry Boessenkool, berry-b@gmx.de, 2011-2014. I'd be interested in hearing what you used the function for.
References
http://uxblog.idvsolutions.com/2011/10/telling-truth.html, https://www.theusrus.de/blog/the-good-the-bad-22012/
See Also
classify, colPointsLegend, colPointsHist
Examples
i <- c( 22, 40, 48, 60, 80, 70, 70, 63, 55, 48, 45, 40, 30, 32)
j <- c( 5, 10, 15, 20, 12, 30, 45, 40, 30, 36, 56, 33, 45, 23)
k <- c(175, 168, 163, 132, 120, 117, 110, 130, 131, 160, 105, 174, 190, 183)
# basic usage:
colPoints(i,j,k, cex=1.5, pch="+", add=FALSE)
# with custom Range:
colPoints(i,j,k, cex=1.5, pch="+", add=FALSE, Range=c(150,190), density=FALSE)
# can be used to allow comparison between several plots
# points outside the range are plotted with col2
# with custom colors:
mycols <- colorRampPalette(c("blue","yellow","red"))(50)
colPoints(i,j,k, cex=1.5, pch="+", add=FALSE, col=mycols)
# With legend title:
colPoints(i,j,k, cex=2, add=FALSE, zlab="Elevation [m above NN.]",
legargs=list(density=FALSE))
?colPointsLegend # to see which arguments can be set via legargs
# colPoints with matrix:
colPoints(z=volcano, add=FALSE)
# image and contour by default transpose and reverse the matrix!
# colPoints shows what is really in the data.
# add single newly measured points to image (fictional data):
mx <- c( 22, 40, 45, 30, 30, 10)
my <- c( 5, 33, 56, 70, 45, 45)
mz <- c(110, 184, 127, 133, 170, 114)
colPoints(mx,my,mz, cex=5, pch="*", Range=c(94, 195), col=seqPal(), col2=NA, legend=FALSE)
points(mx,my, cex=4)
text(mx,my,mz, adj=-0.5, font=2)
# with logarithmic color scale:
shp <- seq(0.2,3, by=0.1)
scl <- seq(0.2,3, by=0.1)
wsim <- sapply(shp, function(h) sapply(scl, function(c) mean(rweibull(1e3, shape=h, scale=c))))
colPoints(shp, scl, (wsim), add=FALSE, asp=1)
colPoints(shp, scl, (wsim), add=FALSE, asp=1, method="log")
# with lines (nint to change number of linear interpolation points):
colPoints(i,j,k, cex=1.5, add=FALSE, lines=TRUE, nint=10, lwd=2)
# With NAs separating lines:
tfile <- system.file("extdata/rivers.txt", package="berryFunctions")
rivers <- read.table(tfile, header=TRUE, dec=",")
colPoints(x,y,n, data=rivers, add=FALSE, lines=TRUE)
colPoints(x,y,n, data=rivers, add=FALSE, lines=TRUE, pch=3, lwd=3)
colPoints(x,y,n, data=rivers, add=FALSE, lines=TRUE, pch=3, lwd=3, nint=2)
colPoints("x","y","n", data=rivers, add=FALSE)
# different classification methods:
# see ?classify
colPoints(i,j,k, add=FALSE) # use classify separately:
text(i,j+1,k, col=divPal(100,rev=TRUE)[classify(k)$index], cex=1)
# Add histogram:
cp <- colPoints(i,j,k, add=FALSE)
do.call(colPointsHist, cp[c("z","at","labels","bb","nbins")])
do.call(colPointsHist, owa(cp[c("z","at","labels","bb","nbins")],
list(bg=5, breaks=5)))
do.call(colPointsHist, owa(cp[c("z","at","labels","bb","nbins")],
list(mar=c(0,0,0,0), x1=0.5, x2=1, y1=0.8,
y2=0.99, yaxt="n")))
# histogram in lower panel:
layout(matrix(1:2), heights=c(8,4) )
colPoints(i,j,k, add=FALSE, y1=0.8, y2=1)
colPointsHist(z=k, x1=0.05, x2=1, y1=0, y2=0.4, mar=3, outer=TRUE)
layout(1)
# Customizing the legend :
cp <- colPoints(i,j,k, legend=FALSE, add=FALSE)
colPointsLegend(x1=0.2, x2=0.95, y1=0.50, y2=0.40, z=k, labelpos=5, atminmax=TRUE, bg=7)
colPointsLegend(x1=0.5, x2=0.90, y1=0.28, y2=0.18, z=k, Range=c(80, 200), nbins=12, font=3)
colPointsLegend(x1=0.1, x2=0.40, y1=0.15, y2=0.05, z=k, labelpos=5, lines=FALSE, title="")
colPointsLegend(z=k, horizontal=FALSE)
colPointsLegend(x1=0.01, y2=0.80, z=k, horizontal=FALSE, labelpos=4, cex=1.2)
colPointsLegend(x1=0.23, y2=0.95, z=k, horizontal=FALSE, labelpos=5, cex=0.8,
dens=FALSE, title="", at=c(130,150,170), labels=c("y","rr","Be"), lines=FALSE)
# For method other than colPoints' default, it is easiest to include these
# options as a list in legargs, but you can also use the invisible output
# from colPoints for later calls to colPointsLegend
do.call(colPointsLegend, cp)
do.call(colPointsLegend, owa(cp, list(colors=divPal(100), cex=1.2)))
# santiago.begueria.es/2010/10/generating-spatially-correlated-random-fields-with-r
if(require(gstat)){
xyz <- gstat(formula=z~1, locations=~x+y, dummy=TRUE, beta=1,
model=vgm(psill=0.025,model="Exp",range=5), nmax=20)
xyz <- predict(xyz, newdata=data.frame(x=runif(200, 20,40),y=runif(200, 50,70)), nsim=1)
head(xyz)
colPoints(x,y,sim1, data=xyz, add=FALSE)
}