Is it possible in ggbiplot package in R to mark some special data points like data point with Alcohol = 13.2 in another color then in the clusters. I want to highlight some specific data points.
data(wine)
wine.pca <- prcomp(wine, scale. = TRUE)
print(ggbiplot(wine.pca, obs.scale = 1, var.scale = 1, groups = wine.class,
ellipse = TRUE, circle = TRUE))
Yes, because ggbiplot is a valid ggplot2 object, you can add geom on it just as you would with other ggplots.
cond <- wine.pca$x[which(wine$Alcohol == 13.2),]
cond[1:2]
Returns:
PC1 PC2
-2.2032498 -0.3324551
Here's the code. Notice we use geom_point and point to PC1 and PC2 respectively for the x and y axes where wine$Alcohol==13.2:
library(ggbiplot)
data(wine)
wine.pca <- prcomp(wine, scale. = TRUE)
ggbiplot(wine.pca, obs.scale = 1, var.scale = 1, groups = wine.class, ellipse = TRUE, circle = TRUE) +
geom_point(aes(x=cond[1], y=cond[2]), col="goldenrod2", size=3)
I am not sure if this can be done within the ggbiplot call, but it is certainly possible by adding another layer. Here is an example:
library(ggbiplot)
data(wine)
wine.pca <- prcomp(wine, scale. = TRUE)
p <- ggbiplot(wine.pca, obs.scale = 1, var.scale = 1, groups = wine.class,
ellipse = TRUE, circle = TRUE)
now subset the points you would like to highlight:
wine.pca$x[wine$Alcohol == 13.2, 1:2, drop = FALSE] #drop = FALSE to keep matrix class
and provide that as data to geom_point:
p + geom_point(data = wine.pca$x[wine$Alcohol == 13.2, 1:2, drop = FALSE],
aes(x = PC1, y = PC2), color = "black")
you can also do something like this:
p + geom_point(data = wine.pca$x[wine$Alcohol == 13.2, 1:2, drop = FALSE],
aes(x = PC1, y = PC2), color = "black", shape = 21, size = 7)
however this might be too much if you want to highlight many points:
p + geom_point(data = wine.pca$x[wine$Alcohol >= 13.2, 1:2],
aes(x = PC1, y = PC2), color = "black", shape = 21, size = 7)
