Skip to content

Map legend

2 messages · Mihai Nica, Roger Bivand

Original
Mihai Nica
# 
Greetings:

I am unable to change the color to some sort of a gray scale in a map. Also, 
I have a hard time adding a meaningful legend to it. These must be ?for 
dummies? question. Looked it up everywhere, but could not find a solution 
(not even a similar question). The map is:


a=read.shape("Density.shp", dbf.data=TRUE, verbose=TRUE)
plot.Map(a, auxvar=pidg$AVG_IGR, xlab="AVG_IGR quintiles", ylab=" ",ol="NA", 
color="blue", nclass=4)

if I change:

color=?gray? or ?black?

Error in hsv(rep(color.list$hsvcol[cind], nclass), c(pr, 1)) :
        bad hsv to rgb color conversion

I am running last R in XP updated.

If anybody has a moment, I appreciate it a lot,


Mihai Nica

_________________________________________________________________
Don?t just search. Find. Check out the new MSN Search!
Roger Bivand
#
On Thu, 16 Feb 2006, Mihai Nica wrote:

            

      
      
      
    

        
It is much easier to use plot() methods for SpatialPolygon objects, and
those methods will be maintained, but plot.Map() isn't going anywhere - 
it tries to do too much with the wrong kinds of objects. 

The help page even explains that you can pass a vector of arbitrary
colours through the fg= argument. The reason for the color argument not
working is that only c("blue", "green", "yellow", "red") are allowed in
its internal color.ramp() function - it is hardcoded, and unnecessary now
that alternatives are available.

Here are the problems:

library(maptools)
x <- read.shape(system.file("shapes/sids.shp", package="maptools")[1])
plot.Map(x, auxvar=x$att.data$BIR74, ol=NA, color="blue", nclass=4) 
plot.Map(x, auxvar=x$att.data$BIR74, ol=NA, color="grey", nclass=4) 

because "grey" isn't in the hardwired list. But using the wonderful
colorRampPalette() function, rolling your own colours is easy, here 100
for one per county:

grays <- colorRampPalette(c("grey90", "grey10"))(100)
plot.Map(x, fg=grays[rank(x$att.data$BIR74)], ol=NA)

or four classes:

moregrays <- colorRampPalette(c("grey90", "grey10"))(4)
brks <- quantile(x$att.data$BIR74)
plot.Map(x, fg=moregrays[findInterval(x$att.data$BIR74, brks, 
  all.inside=TRUE)], ol=NA)

All of the maintained plot() methods - for old-style polylists and new 
style SpatialPolygons use external breaks and colours, because user needs 
differ and advances in R itself (like colorRampPalette()) mean that 
hardwiring things is not a robust option.

If maps need to be re-created often from the same arguments, a custom
function (placing the legend, choosing label sizes and placings, etc.) is
thw way to go using the ingredients above - a custom palette, custom
breaks - for instance using the same breaks on many different data sets,
and findInterval() for continuous-valued variables, and (with care)
as.integer() of factors for categorical variables.

Hope this helps - I think plot.Map will get a "deprecated" warning fairly 
soon!