panel.hexbin(x, y, subscripts, hexbin.data, lwd=add.line$lwd,
lty=add.line$lty, col=add.line$col, ..., lmline=F, text=!fill,
fill=T, maxcount=max(c(0, bin$count)), mincount=1,
col.regions=trellis.par.get("regions")$col, count.to.col, border=F)
panel.hexbin.lmline(x, y, subscripts, hexbin.data, lwd = add.line$lwd,
lty = add.line$lty, col = add.line$col, ...)
panel.hexbin.loess(x, y, subscripts, hexbin.data, lwd = add.line$lwd,
lty = add.line$lty, col = add.line$col, ..., span = 2/3, degree = 1,
family = c("symmetric", "gaussian"), evaluation = 50)
hexbin.data[subscripts,][,c("x","y")]
x and
y observations originate.
hexbin, as produced by the
hexbin
function.
bin=hexbin.data[subscripts, , drop = F] is the part of
hexbin.data that belongs in the current panel.
The bin's
x and
y components provide the
"center of gravity" of the points in each cell (and are the
x
and
y arguments to
panel.hexbin).
cell2xy(bin) provides the the geometric centers of the cells.
bin$count gives the counts in each cell.
lmline is set to
TRUE.
lmline is set to
TRUE.
lmline is set to
TRUE.
TRUE, print the count in each cell.
FALSE, do not color in the hexagonal cells.
count.to.col are subscripts into this list
of colors.
TRUE, draws borders around each hexagonal cell.
panel.hexbin.loess.
1 is locally-linear fitting and
2 is locally-quadratic fitting.
"gaussian" or
"symmetric". In the first case,
local-fitting methods are used. In the second case, the default, local
fitting is used together with a robustness feature that guards against
distortion by outliers.
x and
y
values.
The
x and
y components are required by the
underlying Trellis function, but for this function, they are supplied
by
hexbin.data.
library(bigdata)
xyplot(data=as.bdFrame(fuel.frame), Disp.~Weight|Type,
panel = function(...) { panel.hexbin(...)})
#as.bdFrame requires loadin the big data library.