How to facet data using 3 conditions using ggplot in R?

Question

I have a big data frame (more than 400,000 rows) named as df.

About Dataset

This data set is related to vehicle movements in every 0.1 seconds. The relevant variables for this question are explained below:
class & pclass = Class of vehicle i.e. 1=motorcycle, 2=car, 3=truck
id = Unique ID of a vehicle
frame = Unique ID of a frame in which vehicle was observed. Every frame is 0.1 seconds long
svel = Velocity of a subject vehicle
pvel = Velocity of a vehicle which is in front of (preceding) subject vehicle

Question

I want to do following with the data set:

1. Subset the data set based on the class & pclass. This will create 4 cases: car following car, car following truck, truck following car and truck following truck
2. Create plots for each id in each of the above 4 cases which should be between frame and Velocity (both svel and pvel i.e. 2 lines on a single plot)

What I've tried

df <- data.frame(id=rep(c(1,2,3,4,5,6,7,8,9,10),each=5),
                 frame=rep(1:5,5),
                 class=rep(c(2,3,3,3,2,2,3,2,2,2), each=5),
                 svel=c(15,20,30,15,25,69,45,25,36,45,25,45,45,45,44,40,38,39,39,40,33,34,35,26,50,50,50,50,45,44,43,46,40,35,34,33,32,31,30,32,34,36,38,42,44,46,48,50,52,56),
                 pclass=rep(c(0,2,3,3,3,2,2,3,2,2), each=5),
              pvel=c(0,0,0,0,0,15,20,30,15,25,69,45,25,36,45,25,45,45,45,44,40,38,39,39,40, 33,34,35,26,50,50,50,50,45,44,43,46,40,35,34,33,32,31,30,32,34,36,38,42,44))

I wrote 2 pieces of code to create plots as follows:

  ggplot(data=df, aes(group=id)) + 
      geom_line(mapping=aes(x=frame, y=svel, linetype='subject vehicle')) +
      geom_line(mapping=aes(x=frame, y=pvel, linetype='preceding vehicle')) +
      scale_linetype(name = "Vehicle") +
      facet_grid(class~pclass)

and:

  ggplot(data=df, aes(color=as.factor(id))) + 
      geom_line(mapping=aes(x=frame, y=svel)) +
      geom_point(mapping=aes(x=frame, y=pvel)) +
      scale_linetype(name = "Vehicle") +
      facet_grid(class~pclass)

They somewhat solve the problem of faceting but I want to create a separate plot for every vehicle ID. You can see that right now each plot contains more than 1 id. How can I do that?

Answer 1

Probably you want a list of plots. I like to use plyr::dlply for this. Wrap the code to create a plot in a function, call it something like makePlot. (Don't worry about group = id in your function, plyr will only pass data with 1 id at a time.)

Then you can do this:

library(plyr)
myplots <- dlply(.data = df, .variables = id, .fun = makePlot)

Then, myplots should be a list of ggplots. You can print them one at a time with print(myplots[[1]]), you can continue to modify them, and if you want to arrange them all, this question should at least get you a very good start.

If, on the other hand, you like the faceting approach and would like to use it for id, you can give facet_grid a bigger formula, like facet_grid(class ~ id + pclass), but this could make for a very big plot if you have a lot of ids.

Edits:

I don't know offhand what's wrong with dlply and your makePlot. You can always do a loop instead:

myplots <- list()
for (i in unique(df$id)) {
    myplots[[i]] <- makePlot(filter(df, id == i))
}

Looking more carefully at your function and data, you could restructure things like this:

library(reshape2)
dfm <- melt(df, id.vars = c("id", "frame", "class", "pclass"),
            variable.name = "Vehicle", value.name = "Velocity")
levels(dfm$Vehicle) <- c("Subject Vehicle", "Preceding Vehicle")

Then a plot for an individual id gets as simple as

ggplot(data=subset(dfm, id == 1), aes(x = frame, y = Velocity)) +
       geom_line(mapping=aes(linetype = Vehicle))