Add `sparsity()` function to calculate sparsity of object

[EmilHvitfeldt]

would most likely be applied to tibble, but having matrix and Matrix support wouldn't hurt

some things to think about:

• how should be handle non-numeric values the calculation?
• for sparse columns this becomes easy, but it could be quite slow to implement for longer dense vectors. might have to C it.
• one way we could speed things up would be a sampling/heading of the data. say 100-1000 observations, calculate he sparsity of that. would be worth it at least as an option