Interaction Plot in ggplot2

问题:

I'm trying to make interaction plot with ggplot2. My code is below:

library(ggplot2)
p <- qplot(as.factor(dose), len, data=ToothGrowth, geom = "boxplot", color = supp) + theme_bw()
p <- p + labs(x="Dose", y="Response")
p <- p + stat_summary(fun.y = mean, geom = "point", color = "blue")
p <- p + stat_summary(fun.y = mean, geom = "line", aes(group = 1))
p <- p  + opts(axis.title.x = theme_text(size = 12, hjust = 0.54, vjust = 0))
p <- p  + opts(axis.title.y = theme_text(size = 12, angle = 90,  vjust = 0.25))
print(p)

How can I plot dose-supp level combination means rather than only dose level means which I'm getting here? Thanks in advance for your help.

回答1:

You can precalculate the values in their own data frame:

toothInt <- ddply(ToothGrowth,.(dose,supp),summarise, val = mean(len))

ggplot(ToothGrowth, aes(x = factor(dose), y = len, colour = supp)) + 
    geom_boxplot() + 
    geom_point(data = toothInt, aes(y = val)) +
    geom_line(data = toothInt, aes(y = val, group = supp)) + 
    theme_bw()

Note that using ggplot rather than qplot makes the graph construction a lot clearer for more complex plots like these (IMHO).

回答2:

You can compute your summaries by the appropriate groups (supp):

p <- qplot(as.factor(dose), len, data=ToothGrowth, geom = "boxplot", color = supp) + theme_bw()
p <- p + labs(x="Dose", y="Response")
p <- p + stat_summary(fun.y = mean, geom = "point", color = "blue", aes(group=supp))
p <- p + stat_summary(fun.y = mean, geom = "line", aes(group = supp))
p <- p  + opts(axis.title.x = theme_text(size = 12, hjust = 0.54, vjust = 0))
p <- p  + opts(axis.title.y = theme_text(size = 12, angle = 90,  vjust = 0.25))
print(p)

Or converting to ggplot syntax (and combining into one expression)

ggplot(ToothGrowth, aes(as.factor(dose), len, colour=supp)) +
  geom_boxplot() +
  stat_summary(aes(group=supp), fun.y = mean, geom="point", colour="blue") +
  stat_summary(aes(group=supp), fun.y = mean, geom="line") +
  scale_x_discrete("Dose") +
  scale_y_continuous("Response") +
  theme_bw() +
  opts(axis.title.x = theme_text(size = 12, hjust = 0.54, vjust = 0),
    axis.title.y = theme_text(size = 12, angle = 90,  vjust = 0.25))

---

EDIT:

To make this work with 0.9.3, it effectively becomes Joran's answer.

library("plyr")
summ <- ddply(ToothGrowth, .(supp, dose), summarise, len = mean(len))

ggplot(ToothGrowth, aes(as.factor(dose), len, colour=supp)) +
  geom_boxplot() +
  geom_point(data = summ, aes(group=supp), colour="blue", 
             position = position_dodge(width=0.75)) +
  geom_line(data = summ, aes(group=supp), 
            position = position_dodge(width=0.75)) +
  scale_x_discrete("Dose") +
  scale_y_continuous("Response") +
  theme_bw() +
  theme(axis.title.x = element_text(size = 12, hjust = 0.54, vjust = 0),
        axis.title.y = element_text(size = 12, angle = 90,  vjust = 0.25))

回答3:

If you think you might need a more general approach, you could try function rxnNorm in package HandyStuff (github.com/bryanhanson/HandyStuff). Disclaimer: I'm the author. Disclaimer #2: the box plot option doesn't quite work right, but all the other options are fine.

Here's an example using the ToothGrowth data:

p <- rxnNorm(data = ToothGrowth, res = "len", fac1 = "dose", fac2 = "supp", freckles = TRUE, method = "iqr", fac2cols = c("red", "green"))
print(p)

回答4:

a much easier way. without ddply. directly with ggplot2.

ggplot(ToothGrowth, aes(x = factor(dose) , y=len , group = supp, color = supp)) + 
  geom_boxplot() +
  geom_smooth(method = lm, se=F) +
  xlab("dose") +
  ylab("len")