So in this post:
https://stats.stackexchange.com/a/16335/78575
The author explains a difference of scaling vs not scaling PCA. The two results differ quite a bit on the same data.
Which one then is more accurate and how to infer the accuracy and the need for scaling or not?