For our example here, we'll download US States. (It also includes "state equivalents", e.g., Puerto Rico, Guam, that we'll want to filter out later.) The year shouldn't matter for things like states or counties. It's relevant for other datasets like state legislative districts.