Love the little videos Nature puts out, you can find a bunch more on their YouTube channel. I'd be interested in seeing a similar map based on some measure of semantic/vocabulary similarity rather than just citations. There are probably a lot of connections between papers that aren't captured by citations