How to plot individual components in horizontal bar chart?

Question

I am trying to analyse soccer data, where particular passes and goals are tracked over 3 periods of play or Term. The type of defensive structure or Mode employed by one team is also tracked. An example of my dataset is below:

# Example data
Time <- c(1, 1, 2, 3, 4, 5, 6, 9, 1, 2, 3, 5, 5, 7, 9, 1, 2, 4, 7, 9)
Event <- c("Pass", "Pass", "Pass", "Goal", "Pass", "Pass", "Goal", "Pass", "Pass", 
           "Pass", "Pass", "Pass", "Pass", "Pass", "Pass", "Goal", "Pass", "Pass", "Pass", "Goal")
Term <- c(1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3) 
Symbol <- c("P", "P", "P", "G", "P", "P", "G", "P", "P", 
            "P", "P", "P", "P", "P", "P", "G", "P", "P", "P", "G")
By <- c("Home", "Away", "Home", "Home", "Home", "Away", "Away", "Away", "Home", 
            "Home", "Home", "Away", "Away", "Away", "Home", "Home", "Home", "Away", "Away", "Away")
Mode <- c("Press", "Press", "Press", "Forward", "Forward", "Forward", "Forward", "Press", "Press", 
            "Press", "Press", "Press", "Press", "Forward", "Forward", "Forward", "Forward", "Press", "Press", "Forward")
# Make data.frame
GameData <- data.frame(Time, Event, Term, Symbol, By, Mode)
# Make factors
GameData$Event <- as.factor(GameData$Event)
GameData$Symbol <- as.factor(GameData$Symbol)
GameData$Mode <- as.factor(GameData$Mode)
GameData$Term <- as.factor(GameData$Term)
GameData$By <- as.factor(GameData$By)

I want to visualise when these passes and goals are performed over Time, according to when the Mode is changed. However, when I plot this as a horizontal bar figure in ggplot2, the time is instead summed rather than colour changing at the appropriate time. For example, the maximum time for each term is 9 minutes but the x-axis goes up to 30? My code for the plot is below:

# Load package
require(ggplot2)
# Plot
ggplot(GameData, aes(x = Term, y = Time, fill = Mode)) +
  geom_bar(stat = "identity") +
  geom_text(data = GameData, aes(label = Symbol, colour = By), size = 9) +
  scale_color_manual(values =c("black", "red")) +
  coord_flip() +
  scale_x_discrete(limits = rev(levels(GameData$Term))) 

I feel like this is a silly mistake but where am I going wrong, so the colour/ Mode changes at the corresponding time?

In summary, I want the background of the plot below to have a different colour for each Mode over Time for each Term.

# Plot
ggplot(GameData, aes(x = Term, y = Time, fill = Mode)) +
  geom_text(data = GameData, aes(label = Symbol, colour = By), size = 9) +
  scale_color_manual(values =c("black", "red")) +
  coord_flip() +
  scale_x_discrete(limits = rev(levels(GameData$Term))) +
  theme_classic()

Thanks.

Answer 1

Here is a partial solution using geom_tile() to draw the colored bars (Thanks to @Gregor for the idea).

# Work around to make sure Time=8 appears on x-axis.
GameData$discrete_time = factor(GameData$Time, levels=paste(1:9))

plot1 = ggplot(GameData, aes(y=Term, x=discrete_time)) +
        geom_tile(aes(fill=Mode)) +
        geom_text(aes(label=Symbol, colour=By), size=9) +
        scale_color_manual(values=c("black", "white")) +
        scale_fill_brewer(palette="Set1") +
        scale_x_discrete(drop=FALSE)

ggsave("plot.png", plot=plot1, width=9, height=3, dpi=150)

---

Comments:

• You can fill in the empty tiles by adding appropriate rows to your data (set Symbol to NA to color the tile but leave Pass/Goal blank).
• There are some locations (Term=1, Time=1, for example) where there are two 'Passes' that are fully overlapping. I tried various dodging, jittering, grouping strategies but nothing looked good and revealed the overlapping events. Solving this issue will become more important if you scale up to larger datasets.