People have done this, yes. Often part as larger system tracking tool. Mrtg for instance, or its successor (cricket?). HP Openview in combination with Crystal Reports does more than you want to know. And then there are a billion system administrators each rolling their own. No doubt some of it was written in Perl.

I've used mrtg, HP Openview, Big brother, and a handful of solution written by myself to track disk (or rather, volume) usage. Some of it written in Perl. But also in C and shell/awk.

But none of it was based on File::Find, or on find(1). Summing the sizes of files reported this way doesn't give you accurate information - hard links are counted twice, sparse files are counted for way too much, open, but deleted files are not counted at all, it's too easy to forget counting the size of a directory, and special files are a problem. And 1000 files of one character take a lot more space than a single file containing 1000 characters. Just ask your OS, or your volume manager how space is in use. And don't forget to record your inode usage - data blocks aren't the only thing that can cause a filesystem to fill up.