Matrix normalization over multiple runs, what does this code do?

Question

I have several numpy matrices collected over some time. I now want to visualize these matrices and explore visual similarities among them. The matrices contain small numbers