Use hexdump -C for more (imho) readable output. If you're an emacs person, check out hexl-mode, for read-write version with a similar layout.

Also, it looks like the data you are scraping is unicode. &#65533 appears to be this character.

If you're dealing with unicode data, you may want to be a bit more careful and convert characters rather than stripping them out, but it depends on your environment and application.