ggplot geom_boxplot and stat_smooth

Hi Petr:

I had to do a little bit of finagling, but this seems to work. I 

did the following:

(i) coordinated the dodging of the points and boxplots by typ within 

(ii) summarized the group medians in a separate data frame and added an
additional
    column to compensate for the offsets in the boxplots due to dodging
(iii) used geom_line() with group typ to get the connecting lines.

# library(ggplot2)

# Summarize group medians
ads <- ddply(ad, .(konc.f, typ), summarise, m = median(bel))
# Offset x-positions
ads$pos <- rep(c(1:4, 6, 7), each = 4) + rep(c(-0.3, -0.1, 0.1, 0.3), 6)

# Note use of position_dodge() with same width in both the boxplot and 

geoms
# Add geom_line with summarized data frame and offset x-positions
# group = typ produces a separate line per typ

p + geom_boxplot(position = position_dodge(width = 0.8)) +
    geom_point(position = position_dodge(width = 0.8)) +
    geom_line(data = ads, aes(x = pos, y = m, colour = typ, group = 

I chose to create a separate data frame for the summaries rather than 

stat_summary(), for example, because it is a bit easier to plot 

layers that way.

HTH,
Dennis

On Mon, Jan 24, 2011 at 11:05 PM, Petr PIKAL <petr.pikal at precheza.cz> 

Dear all

I would like to superpose some smoothing line through boxplot in 

dput(ad)

structure(list(konc.f = structure(c(1L, 1L, 1L, 1L, 2L, 2L, 2L,
2L, 3L, 3L, 3L, 3L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 6L, 6L, 6L,
7L, 7L, 7L, 7L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 3L, 3L, 3L, 3L,
4L, 4L, 4L, 4L, 4L, 4L, 4L, 6L, 6L, 6L, 7L, 7L, 7L, 1L, 1L, 1L,
1L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 6L,
6L, 6L, 6L, 7L, 7L, 7L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 3L, 3L, 3L,
3L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 6L, 6L, 6L, 6L, 7L, 7L, 7L), .Label =
c("(189,196]",
"(196,202]", "(202,208]", "(208,215]", "(215,221]", "(221,227]",
"(227,234]"), class = "factor"), bel = c(99.28, 99.29, 99.25,
98.13, 99.51, 99.21, 99.09, 97.84, 98.97, 98.48, 98.64, 98.09,
98.44, 98.19, 98.25, 97.54, 99.11, 98.23, 97.62, 97.01, 97.62,
97.58, 97.42, 97.75, 97.16, 96.79, 96.82, 98.8, 99.02, 98.86,
98.85, 99.25, 98.46, 98.68, 98.3, 98.67, 98.54, 98.39, 98.18,
98.54, 99.13, 98.92, 98.29, 98.78, 98.58, 98.78, 98.9, 98.18,
97.5, 98.63, 97.53, 96.55, 96.52, 96.23, 95.54, 95.54, 96.33,
95.91, 95.41, 94.98, 94.69, 93.95, 94.05, 95.22, 94.6, 94.27,
93.44, 95.15, 94.62, 93.86, 94.51, 94.83, 93.66, 92.95, 94.4,
93.17, 92.77, 95.79, 95.03, 94.96, 95.94, 95.71, 95.19, 95.11,
94.91, 94.42, 94.68, 94.91, 94.66, 94.05, 93.57, 93.43, 94.77,
93.58, 93.84, 93.24, 94.45, 93.57, 93.46, 92.38, 92.39, 94.07
), typ = structure(c(3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L,
3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L,
3L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L,
4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 4L, 1L, 1L, 1L, 1L, 1L, 1L,
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L,
1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L,
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L), .Label = c("alrut",
"nealrut", "stan", "vlakan"), class = "factor")), .Names = c("konc.f",
"bel", "typ"), row.names = c(NA, -102L), class = "data.frame")

Here is what I did

p<-ggplot(ad, aes(x=konc.f, y=bel, colour=typ))

geom_smooth: Only one unique x value each group.Maybe you want 

1)?

I get nice picture with boxes but I expected to get something like
smoothing line through box centres.

Is it possible without some hack to stat_smooth code?

Regards
Petr

   [[alternative HTML version deleted]]