A.C, I would be conservative in answering your question. It is easy to create a category, but the category then is only useful if (1) it aggregates a group of people that is helpful for genealogy, and (2) it is used by enough other people so that the category will contain profiles worth finding.
So I would say that SOME Censuses would not be useful at all. The 1940 United States Census is (and prior censuses are) very useful to go and find data on but having everyone who appears in the 1940 Census categorized that way would not, in my opinion, add any information that could not be gotten easily by going to the census itself. Even the 1776 Maryland census, which is much more limited and affects fewer numbers -- yes, the data is very important to capture in the biography with appropriate inline sourcing, but categorizing this person according to the census would just be duplicative. Secondly, if the category indeed IS useful, it would depend on getting all the eligible profiles categorized, which would require the work of a lot of people. So it would seem to me to be a big expenditure of effort for limited results.
Now I could see the value in what I call a "micro-category" which is created by a particular researcher for a particular purpose, e.g. grouping 25 people who are all on the same street in the 1870 Census. You are wondering what their interconnections are, so you create this micro-category and link the 25 profiles to it. Now you have a quick way to go back and forth among the 25. You may add some common features you discover to the narrative lines in the category. The value of this microcategroy is not limited only to you, but it really is limited only to anyone interested in those 25 people. There would be no expectation that all the streets in the 1870 Census would receive similar detailed attention, although you would properly link this microcategory to the town in which the street appears and other appropriate higher categories so that an interested person could find it..
As a member of the categorization project, I always try to ask (and answer) the question, "how does this category contribute to genealogical research and understanding?"