Can I manipulate aggregate tables at the census tract level? In essence, could I create dummy variables, such as: white, black, hipanic, other for 100 census tracts? Or is this only possible with individual level data?