I left out one crucial piece of information. The legend should show
all possible values because I am producing the same map for different
countries. The solution I found was to add observations for
non-existing regions that contain all possible values of the mapped
variable. The code below shows how this works with the example by
Scott.
use "italy-regionsdata.dta", clear
keep id relig1
gen x = id*20
xtile key = relig1, cutpoints(x)
replace key = 6 if id >18
* Add observations with all 6 values of the mapped variable
count
local oldobs = r(N)
local newobs = r(N) + 6
sum id
local idmax = r(max)
set obs `newobs'
forvalues i = 1/6 {
local obs = `oldobs' + `i'
* Create IDs for non-existing regions on the map
replace id = `idmax' + `i' in `obs'
replace key = `i' in `obs'
}
spmap key using "Italy-RegionsCoordinates.dta", id(id) ///
clmethod(unique) fcolor( "215 25 28" "253 174 97" ///
"255 255 191" "171 217 233" "44 123 182" "225 225 225" ) ///
name(gr1,replace)
Friedrich
On 9/6/07, Scott Merryman <[email protected]> wrote:
> I am not sure I understand.
>
> The variable "key" could take on 6 values, but in this particular
> realization you have 4 values = 2, 3, 4, 6
>
> But you want -spmap- to only use the 2nd, 3rd, 4th, and 6th values
> from your predfined -fcolor()- list. Is this correct?
>
> For example:
>
> use "Italy-RegionsData.dta", clear
> keep id relig1
> gen x = id*20
> xtile key = relig1, cutpoints(x)
> replace key = 6 if id >18
>
> spmap key using "Italy-RegionsCoordinates.dta", id(id) ///
> clmethod(unique) fcolor( "215 25 28" "253 174 97" ///
> "255 255 191" "171 217 233" "44 123 182" "225 225 225" ) ///
> name(gr1,replace)
>
> This would produce a map with the "wrong colors." If you were doing
> by hand you could eliminate the colors that do not match the "key"
>
> spmap key using "Italy-RegionsCoordinates.dta", id(id) ///
> clmethod(unique) fcolor( "253 174 97" ///
> "255 255 191" "171 217 233" "225 225 225" ) ///
> name(gr1,replace)
>
> You could automate this by looping through your "key" values to match
> them with your predefined -fcolor()- list:
>
> levelsof key, local(levels)
> local a "215 25 28" "253 174 97" "255 255 191" "171 217 233" "44 123
> 182" "225 225 225"
> foreach l of local levels {
> local b: word `l' of "`a'"
> local colorlist "`colorlist' "`b'" "
> }
>
> spmap key using "Italy-RegionsCoordinates.dta", id(id) ///
> clmethod(unique) fcolor(`colorlist') ///
> name(gr2,replace)
>
>
> Scott
>
> On 9/5/07, Friedrich Huebler <[email protected]> wrote:
> > When I draw a choropleth map with -spmap- from SSC, the regions have
> > the wrong colors. I define a variable called "key" with the values 1
> > to 6, recoded from a variable with the range 0 to 100 percent. 1 to 5
> > identify quintiles and 6 identifies regions with missing data. My
> > -spmap- command is shown below, with fill colors specified with RGB
> > values.
> >
> > #delimit ;
> > spmap key using map.dta,
> > id(id)
> > clmethod(unique)
> > fcolor("215 25 28" "253 174 97" "255 255 191" "171 217 233" "44 123
> > 182" "225 225 225")
> > ;
> >
> > None of the regions have a value below 20 and none have a value
> > between 60 and 80, so that the first and fourth color in the list
> > above ("215 25 28" = dark red and "171 217 233" = light blue) should
> > not appear on the map. However, the map is drawn with the first three
> > colors from the fcolor list instead of the second, third and fifth
> > color. All regions with the value 2 are filled with the color "215 25
> > 28", all regions with the value 3 are filled with the color "253 174
> > 97", and all regions with the value 5 are filled with the color "255
> > 255 191".
> >
> > What am I doing wrong?
> >
> > Friedrich
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
