Counting nonzero values in a dynamic range

Hi,
I have several large sparse matrices containing duplicate headers for row and column vectors. Due to this, it is likely that multiple rows with same row header may have different values in a single column vector and vice versa. As an intermediate step, I need to calculate if, and how may, duplicates exist in relation to each unique row header. I can do it for individual rows/columns using, for instance, "=COUNTIF(C$4:C$9,">"&0)" but I'm wondering if it can be done for the whole matrix.

The example data is attached in which lower matrix has to be filled with count of non-zero values corresponding to each unique row header.

I shall be grateful for any help.

Best regards,
guest