I'm interested in comparing home ownership rates between different races/ethnicities. Based on the way the questions are structured, the easiest way would be to have two separate plots, one for race and one for ethnicity. However, I personally find this less compelling. Another option I've seen is to lump everyone who identifies as Hispanic into one group, and then group everyone else by race. However, this erases the potentially significant portion of the population that is multiracial. I'm wondering what are considered current best practices for visualizing race and ethnicity data given how the census currently structures those questions.
I think you’re absolutely right — Hispanic is not a race and we have no business treating it like one just to make everything sum to 100% for a pie chart. I include Hispanics and combinations in each race group (except White alone, non Hispanic which is the definition of non-minority and increasingly called just White) — it works for any visualization except pie charts, though you could use them one at a time (Black alone or in combination, Hispanic or not, vs everybody else)