Now I have this:
trellis.par.set( peelTheme)
plot( spplot( thumbDf, "pri",
main= "Primary cover",
mapasp = "iso",
colorkey= FALSE),
panel.width= list( x=0.6, units= "npc"),
split= c( 1, 1, 3, 1),
position= c( 0, 0.2, 1, 0.8),
more= TRUE)
plot( spplot( thumbDf, "sec",
main= "Secondary cover",
mapasp = "iso",
colorkey= list( space= "right")),
panel.width= list( x=0.6, units= "npc"),
split= c( 2, 1, 3, 1),
position= c( 0, 0.2, 1, 0.8),
more= TRUE)
trellis.par.set( seqTheme)
plot( spplot( thumbDf, "pct",
main= "Classification confidence, %",
mapasp = "iso",
at= seq( 0, 100, by= 5),
cuts= 5),
panel.width= list( x=0.6, units= "npc"),
position= c( 0, 0.2, 1, 0.8),
split= c( 3, 1, 3, 1))
The panels are the same size, but I'm not sure I understand what is going on
and have a suspicion that there must be a better way that is more generally
applicable without having to fiddle with so many constants. It's not clear
that the mapasp= arguments have any effect and I don't really understand the
implication of its default for spplot() for geographic data (which is what I
am plotting) in light of the note in the man page of plot.trellis that for
panel.width and panel.height. It says that "[these options] should not be
used in conjunction with non-default values of the aspect argument in the
original high level call (no error will be produced, but the resulting
behaviour is undefined)."
-----
Neil Best
Computation Institute
University of Chicago