Why the numpy correlation coefficient matrix and the pandas correlation coefficient matrix different when using np.corrcoef(x) and df.corr()?
x = np.array([[0, 2,