R help - Sep 2024 - Getting individual co-ordinate points in k medoids cluster

Hello  I am using k medoids in R to generate sets of clusters for datasets
through time.  I can plot the individual clusters OK but what I cannot find
is a way of pulling out the co-ordinates of the individual points in the
cluster diagrams - none of the kmed$... info sets seems to be this.

Beneath is an example of a k medoid prog using the built in US arrests
dataset - this is not the data I am working with but it illustrates the
issue easily

library(factoextra)
library(cluster)
set.seed(170924)
df <- USArrests
df <- na.omit(df)
df <- scale(df)
kmed <- pam(df, k = 4)
fviz_cluster(kmed, data = df)

Thanks for any help or pointers
Nick Wray

	[[alternative HTML version deleted]]

Hi Nick,

If you're asking about the coordinates used in the plot produced by
fviz_cluster(), you're looking in the wrong place.

The 2D plot isn't anything intrinsic to the clustering. Instead, that
plot is of the first two principal components. From the help:

"Observations are represented by points in the plot, using principal
components if ncol(data) > 2."

The easiest way to get the coordinates is probably to use prcomp
yourself, as fviz_cluster does.
Specifically, it does:

pca <- stats::prcomp(data, scale = FALSE, center = FALSE)

That's also where the percentage values in the axis labels come from:
those are the variance explained by each of the first two PCA axes.

Sarah


On Tue, Sep 17, 2024 at 8:01?AM Nick Wray <nickmwray at gmail.com>
wrote:
>
> Hello  I am using k medoids in R to generate sets of clusters for datasets
> through time.  I can plot the individual clusters OK but what I cannot find
> is a way of pulling out the co-ordinates of the individual points in the
> cluster diagrams - none of the kmed$... info sets seems to be this.
>
> Beneath is an example of a k medoid prog using the built in US arrests
> dataset - this is not the data I am working with but it illustrates the
> issue easily
>
> library(factoextra)
> library(cluster)
> set.seed(170924)
> df <- USArrests
> df <- na.omit(df)
> df <- scale(df)
> kmed <- pam(df, k = 4)
> fviz_cluster(kmed, data = df)
>
> Thanks for any help or pointers
> Nick Wray
>
>         [[alternative HTML version deleted]]
>

-- 
Sarah Goslee (she/her)
http://www.numberwright.com