An equivalent function for scipy.stats.multivariate_normal.rvs for skewed data

Question

I have a dataset which is consists of multiple rows (entries) and columns (features). I want to generate dummy data from the real data I have.

First, I calculated the