k-means analysis: how to convert data into numeric? - r

Question

I want to perform a k-means analysis in R. For that I need numeric data. I tried the following
unlist(pca)
as.numeric(pca)
lapply(pca,as.numeric(pca))
pca is just "normal" Principal Component Analysis data, showed in a plot (with fviz_pca_ind() function).
By the way, when I try to run the k-means analysis, it gives me "list object cannot be coerced to type double". That is why I thought to turn everything into numeric.
How to convert the pca-data into numeric?
Thank you ;)

Answer 1

You're almost correct
lapply(pca,as.numeric)
as.numeric is a function and therefore an object. You need to pass it to lapply() as an object and therefore without the quotation marks.

Answer 2

Most pca objects should return you a list, and you should show which package or function is used to perform the pca, so we can see what's in the list
For example, if you use prcomp, it returns a list of eigenvectors / loadings ($rotation) and principal components ($x). I suppose you are trying to do k-means on the principal componets, and you can do it like:
# perform pca
pca = prcomp(USArrests,scale=TRUE)
# we call out the PCs using pca$x
# and kmeans
kmeans_clus = kmeans(pca$x,3)
## plot
# define colors
COLS = c("#65587f","#f18867","#e85f99")
plot(pca$x[,1:2],col=COLS[kmeans_clus$cluster],pch=20)
legend("topright",fill=COLS,legend=1:3,horiz=TRUE)