I love how PCA results differ widely depending on whether you normalise input data or not, and whether you apply it to the covariance or correlation matrix. I think figuring the correct approach out for this dataset might be a tomorrow problem. If anyone has a great go-to resource, please link me!
