Covariance matrix

Maybe your problem has something in common with this thread:

But you need the data points (as ntuples) to work, rather than histos. Perhaps the experts can give a more relevant answer.

Best regards,