I'd like to aggregate U.S. census tract data, for several variables, to sub-county enumeration units. I want to ensure the demographic composition, by percent, of the aggregate geographies is somewhat similar to the tracts from which they are constructed. Would Similarity Search, specifically SIMINDEX, be useful for this purpose? I'm thinking of using SIMINDEX at census tract level. I'd assign the aggregate value to each census tract that makes up a specific aggregated enumeration unit to compare to the original census tract percent. If I use Attribute profiles, what would be a suitable threshold? Perhaps > 0? Thanks!