Skip to content

color category SpatialGridDataFrame SpatialPolygonsDataFrame Corine Land Cover Global Land Cover 2000

6 messages · Patrick Giraudoux, Renaud Lancelot

#
Dear all,

I am wondering what is the safest way to handle colors when 
SpatialGridDataFrame data are categories. spplot() or image() are 
extremely convenient when attributes are continuous variables or so (eg 
altitude, rainfall, etc...), but the colorRamp() system and the 
col.regions, col arguments, though extremely powerful in other cases, is 
not really adapted to category data such as Corine LandCover or Global 
Land Cover 2000 codes. Thus to link pixels to the right color may be 
quite tedious (however I may have missed something...).

One option I though about was to change raster into shapefile (or 
SpatialGrid into SpatialPolygons objects), but again the straight way to 
get right colors is not evident.

Any idea ?

Patrick
#
Dear Patrick,

I recently met this issue. I finally used image() on the 
SpatialGridDataFrame, defining the appropriate colors using the col 
argument (?tonnant, non ;-)). I have defined the legend using legend(), 
and I used split.screen() to arrange the plots. If you wish, I can send 
you the whole stuff (data and code) in a private message (huge datasets).

Renaud

Patrick Giraudoux a ?crit :

  
    
#
Renaud Lancelot a ?crit :
It was my point. In this case, you must have defined many breakpoints 
(to create appropriate intervals for each code) to get all the codes at 
the right color, which is quite tedious, isn't it ? I wonder if this can 
be simplified. Ideally importing a color table. Maybe a new function to 
write (thus to put on the many pending "to do").
I appreciate the proposal, but I was just interested in some general 
principles and advise...

Have a nice evening...

Cheers,

Patrick
#
Patrick Giraudoux a ?crit :
Indeed, that's what I did (the color table was stored in a dbf file).

Happy new year !

Renaud
#
Patrick Giraudoux a ?crit :
PS: by the way if you connect to http://pagesperso-orange.fr/giraudoux/, 
you can download an unofficial package 'pgirbric' (when I am sure things 
work well I move them to pgirmess). I have written some functions (based 
on splancs and sp) to extract info from rasters using buffer centred on 
points or polygons. A quite common business in my field of ecology but 
with no programmes (to my knowledge) doing it easy. The latest one are 
on the development version (I had quite busy Xmas vacation playing on 
Alevolar echinococcosis distribution in China).

Cheers,

Patrick
#
Patrick Giraudoux a ?crit :
PS: by the way, Renaud, if you connect to 
http://pagesperso-orange.fr/giraudoux/, you can download an unofficial 
package 'pgirbric' (when I am sure things work well I move them to 
pgirmess). I have written some functions (based on splancs and sp) to 
extract info from rasters using buffer centred on points or polygons. A 
quite common business in my field of ecology but with no programmes (to 
my knowledge) doing it easy. The latest functions are on the development 
version pgirbric 1.1.4 (I had quite busy Xmas vacation playing on 
Alevolar echinococcosis distribution in China).

Feed-back and suggestions appreciated if they are usable to your own work.

sp, splancs, maptools, spdep, rgdal, etc... are really great things ! 
GRASS and QGIS also.

Cheers,

Patrick