Could we add a subgraph example?
It would help visualize what variables, entities and attributes etc. are.
Also, for some/each of the subsection (esp. variables), it might help to explain when such node type should used. Again, maybe an example suffices.