plotting the whole data within each facet using facet_wrap and ggplot2

Question

I am trying to plot line graphs for and facet_wrap for each dataset. What I would love to have is in light grey, transparent or something, all datasets in the background.

df <- data.frame(id=rep(letters[1:5], each=10),
                 x=seq(10),
                 y=runif(50))

ggplot(df, aes(x,y, group=id)) + 
  geom_line() + 
  facet_wrap(~ id)

This graph is how far I get, but I would love to have all the other missing 4 lines in each graph as well... In any way I try to use facet_wrap, I get only the data of a single line.

What I would expect is something like this for each facet.

ggplot(df, aes(x,y, group=id)) + 
  geom_line() + 
  geom_line(data=df[1:10,], aes(x,y, group=id), size=5)

Answer 1

Here's another approach:

First add a new column identical to id:

df$id2 <- df$id

Then add another geom_line based on the df without the original id column:

ggplot(df, aes(x,y, group=id)) + 
  geom_line(data=df[,2:4], aes(x=x, y=y, group=id2), colour="grey") +
  geom_line() + 
  facet_wrap(~ id)

Answer 2

Here is an approach. It might not be suitable for larger datasets, as we replicate the data number_of_facets-times.

First, we do some data-wrangling to create this desired dataframe. df$obs_id <- 1:nrow(df) #unique ID for each observation

#new data with unique ID's and 'true' facets
df2 <- expand.grid(true_facet=unique(df$id), obs_id=1:nrow(df))

#merge them
dat <- merge(df,df2,by="obs_id",all=T)

Then, we create a flag defining the 'true' faceted variable, and to discern background from foreground.

dat$col_flag <- dat$true_facet == dat$id

Now, plotting is easy. I've used geom_line twice instead of scales, as that was easier than to try to fix the ordering (would lead to black being plotted below grey).

p1 <- ggplot(dat, aes(x=x,y=y, group=id))+
  geom_line(color="grey")+
  geom_line(dat=dat[dat$col_flag,],size=2,color="black")+
  facet_wrap(~true_facet)