Consider this a follow-up to this question.
I transformed a census block group vector map (decided to use this instead of census tracts) into a raster map and generated a kernel density map using supermarket points. Since I want the average density per census block group, I ran the Zonal Statistics as table tool and specified the value for different CBGs as the zone. Sample output is below.
My cell size is 10x10 and the projection unit is in feet (kernel density area units defaulted to square miles, I'm not sure how much this matters). My issue is that I'm not entirely sure what these numbers represent? Are they the average number of stores per 100 sq ft? And if yes, what number do I need for the cell size so that the mean represents the average density per square mile? I've read other materials trying to figure this out but it still isn't clear to me.