I like your goals! For point (1), as I'm sure you're aware, different types of meat have
different carbon footprints. Perhaps your goal could reflect that? For instance, you could set a carbon budget (say), and therefore beef and lamb (sorry) would be costlier.
Python is also
excellent; it's the language of choice for a lot of deep learning libraries like keras and pytorch. I have found that R is more common in the bioinformatics community. Between the two, I prefer python - and much of the knowledge is transferrable between them, though not the syntax.