If you try limiting the data to just zoom-level-0 blocks, though, you're not going to get a very readable image (in terms of text _and_ in terms of road lines).