Been talking a lot to my girlfriend's brother (a materials scientist) about cool statistics used in materials to guide experimentation & predict materials properties.

It's really cool! It's an extremely "hard" science in that the quantities under study are genuinely real things (unlike the SDF, or risk aversion, for example).

It reminds me of learning statistics, where lots of examples are about real stuff. The t-test, for example, was invented by a chemist/brewer at Guinness!