R help - Nov 2006 - filled.contour and NA's

Hi,

I'm trying to do a filled.contour plot where some points are labelled as 
NA. How do I could plot this kind of graphics, so NA points are coloured 
black,  keeping the levels of remaining points. NA's values represent 
land points (meaningless), and what I want to plot is the levels of a 
variable over the sea.

x<-seq(length=21, from=-10, by=.25)
y<-seq(length=8, from=36, by=.25)

filled.contour(x,y,t(ene))

dput(ene,control="all")
structure(c(0.941176470588235, 0.705882352941177, 0.588235294117647,
0.470588235294118, 0.588235294117647, 0.588235294117647, 0.588235294117647,
0.705882352941177, 0.647058823529412, 0.823529411764706, 0.588235294117647,
0.588235294117647, 0.529411764705882, 0.647058823529412, 0.588235294117647,
0.764705882352941, 0.823529411764706, 0.764705882352941, 0.647058823529412,
0.705882352941177, 0.529411764705882, 0.647058823529412, 0.529411764705882,
0.529411764705882, 0.705882352941177, 0.647058823529412, 0.764705882352941,
0.705882352941177, 0.529411764705882, 0.705882352941177, 0.647058823529412,
0.411764705882353, 0.88235294117647, 0.823529411764706, 0.764705882352941,
0.764705882352941, 0.705882352941177, 0.647058823529412, 0.529411764705882,
0.529411764705882, 0.764705882352941, 0.941176470588235, 0.941176470588235,
0.88235294117647, 0.722222222222222, 0.588235294117647, 0.5625,
0.5, 0.823529411764706, 0.823529411764706, 0.833333333333333,
1, 0.529411764705882, 0.909090909090909, 0.818181818181818, 
1.33333333333333,
0.941176470588235, 0.88235294117647, 0.823529411764706, 0.833333333333333,
0.6, NA, NA, NA, 0.705882352941177, 0.88235294117647, 0.823529411764706,
0.722222222222222, 0.75, NA, NA, NA, 0.647058823529412, 0.705882352941177,
0.764705882352941, 0.722222222222222, 0.5, 1.66666666666667,
NA, NA, 0.529411764705882, 0.647058823529412, 0.647058823529412,
0.588235294117647, 1.16666666666667, 1.5, NA, NA, 0.529411764705882,
0.470588235294118, 0.588235294117647, 0.428571428571429, 1.33333333333333,
1.5, NA, NA, 0.470588235294118, 0.470588235294118, 0.529411764705882,
0.727272727272727, 1.66666666666667, 1.5, NA, NA, 0.470588235294118,
0.555555555555556, 0.588235294117647, 1.16666666666667, 1.66666666666667,
1.66666666666667, NA, NA, 0.611111111111111, 0.529411764705882,
0.583333333333333, 1.16666666666667, 1.16666666666667, NA, NA,
NA, 0.428571428571429, 0.875, 1.33333333333333, 1, 1.33333333333333,
NA, NA, NA, 0.833333333333333, 0.833333333333333, 1.5, 1.33333333333333,
NA, NA, NA, NA, 2.16666666666667, 1.66666666666667, NA, NA, NA,
NA, NA, NA, 1.5, 0.833333333333333, 0.5, NA, NA, NA, NA, NA,
0.666666666666667, 0.666666666666667, 1.16666666666667, NA, NA,
NA, NA, NA, 0.666666666666667, 0.666666666666667, 0.666666666666667,
NA, NA, NA, NA, NA), .Dim = as.integer(c(8, 21)))

On 11/29/06, antonio rodriguez <antonio.raju at gmail.com>
wrote:
> Hi,
>
> I'm trying to do a filled.contour plot where some points are labelled
as
> NA. How do I could plot this kind of graphics, so NA points are coloured
> black,  keeping the levels of remaining points. NA's values represent
> land points (meaningless), and what I want to plot is the levels of a
> variable over the sea.

Hi Antonio,

I haven't seen a reply to your question yet, so I'll make a try.

I'm having the same issue myself. What I've ended up doing is
replacing NA's with a big negative value, define "levels" as one
color
for negative values, and a regular scale above. This gives a fairly
good separation between what areas contain data, and what doesn't. To
make NA's black, I suppose one would have to define your own
colour.palette, but I haven't looked into it.

Hope this helps, and let me know if you've found a better solution!

Best,

Gustaf Rydevik

> I'm trying to do a filled.contour plot where some points are labelled
as
> NA. How do I could plot this kind of graphics, so NA points are coloured
> black,  keeping the levels of remaining points. NA's values represent
> land points (meaningless), and what I want to plot is the levels of a
> variable over the sea.
You can also do this using ggplot, although you don't have quite as
much control over the appearance of the contours (you do get more
control over other things)

install.packages("ggplot", dep=T)
library(ggplot)

# Set up appropriate data structure
dimnames(ene) <- list(y, x)
names(dimnames(ene)) <- c("y", "x")
enem <- melt(ene)

p <- ggplot(na.omit(enem), aes=list(x=x, y=y, z=value, fill=value))
p <- ggtile(p)
p <- ggcontour(p)
scfillgradient(p)
scfillgradient(p, low="white", high="blue")

Regards,

Hadley