I suggest that the event type “Census / Residence” be split into two separate event types. Why? Because a person being enumerated on a census return at a certain place does not necessarily mean they resided there.
Most often people were enumerated at their place of residence, but exceptions to this are not at all uncommon. I can think of many, many exceptions just from my own tree: one great-grandfather is on the 1900 U.S. census in an Army unit in the Philippines, i.e., as part of U.S. forces in the Spanish-American War. Another ancestor of mine closed up her permanent residence after being widowed and traveled between the homes of her several children for the remainder of her life.
The notoriously sloppy 1870 census had an entire second enumeration done in many places. One of my NYC ancestors actually appears four times on the 1870 federal census—on both enumerations he appears once with his wife and children, and once in the residence of his aged mother and stepfather.
Residence events don’t coincide with census years, and can reflect a person’s appearance in any of a range of record types: a deed or other transaction, a probate case, a court record that says John Jacobs of Amesbury gave testimony, etc., etc.
So, currently, my aforementioned Spanish-American War soldier’s enumeration on the 1900 census in the Philippines appears in his timeline as a Residence event, as do all other census enumerations. The distinction between a census enumeration and a residence event is important, and right now it can’t be made in Rootsfinder.