ggplot geom_tile overlay plot with points - r

Question

severity <- c("Major","Serious","Minor","Negligible")
probability <- c("Highly Probable","Probable","Possible","Remote","Unlikely","Impossible")
df <- expand.grid(x=severity,y=probability)
df$x <- factor(df$x, levels=rev(unique(df$x)))
df$y <- factor(df$y, levels=rev(unique(df$y)))
df$color <- c(1,1,2,2,1,2,2,2,2,2,2,3,2,2,3,3,2,3,3,3,3,3,3,3)
ggplot(df,aes(x,y,fill=factor(color)))+
geom_tile(color="black")+
scale_fill_manual(guide="none",values=c("red","yellow","green"))+
scale_x_discrete(expand=c(0,0))+scale_y_discrete(expand=c(0,0))+
labs(x="",y="")
Produces a risk assesssment score card chart. I want to add points by using a csv file by adding a record. Each record has 3 fields, a item name, x, and y coordinate. x= severity and y = probability.
da <- data.frame(list(name=c("ENVIRONMENTAL","COSTS","SUPPLY","HEALTH"),
severity=c("Major","Serious","Minor","Serious"),
probability=c("Probable","Possible","Probable","Unlikely")))
da
name severity probability
1 ENVIRONMENTAL Major Probable
2 COSTS Serious Possible
3 SUPPLY Minor Probable
4 HEALTH Serious Unlikely
> p1 <- p + data.frame(da, aes(severity, probability)) + geom_point()
Error in as.data.frame.default(x[[i]], optional = TRUE, stringsAsFactors = stringsAsFactors) :
cannot coerce class ""uneval"" to a data.frame
>
> d <- data.frame(list(name=c("ENVIRONMENTAL","COSTS","SUPPLY","HEALTH"),
severity=c(2,3,4,1),probability=c(3,5,4,6)))
> d
name severity probability
1 ENVIRONMENTAL 2 3
2 COSTS 3 5
3 SUPPLY 4 4
4 HEALTH 1 6
> ggplot(d,x=severity, y=probability)+ geom_point()
Error in exists(name, envir = env, mode = mode) :
argument "env" is missing, with no default
How can I add points to the ggplot / geom_tile graph?

Answer 1

You can't add a data.frame to a plot (not like that, at least...). What you can do is add a new layer, geom_point(), and specify the data.frame it comes from. To make things work, you should have the columns from any aesthetics you still want to use (here, x and y) have the same names in both data.frames.
# It's better practice to modify your data
# then to convert to factor within the plot
df$color <- factor(c(1,1,2,2,1,2,2,2,2,2,2,3,2,2,3,3,2,3,3,3,3,3,3,3))
# get some meaningful names, that match da and d
names(df)[1:2] <- c("severity", "probability")
p <- ggplot(df, aes(x = severity, y = probability)) +
# moved fill to the geom_tile layer, because it's only used there
geom_tile(color = "black", aes(fill = color)) +
scale_fill_manual(guide = "none", values = c("red", "yellow", "green")) +
scale_x_discrete(expand = c(0, 0)) +
scale_y_discrete(expand = c(0, 0)) +
labs(x = "", y = "")
# alsonoticehowaddingspacesmakesiteasiertoread
# Using the same column names? Yup! Now it's this easy:
p + geom_point(data = da) +
geom_point(data = d, color = "dodgerblue4")