It’s common to truncate US zip codes to the first three digits for privacy reasons. Truncating to the first two digits is less common, but occurs in some data sets.
HIPAA Safe Harbor requires sparse 3-digit zip codes to be suppressed; even when rolled up to three digits some regions are still sparsely populated.
How sparse can a two-digit zip code region be? Empty. The population of all zip codes starting with 09 is zero. That’s because all zip codes of the form 09XXX are overseas military bases and not anyone’s permanent residence.
Aside from that, the least populated 2-digit zip code is 69 with a population of about 175,000.
The most populated 2-digit zip code is 33 with a population of about 10,500,000.
So the ratio of the largest population to the smallest non-zero population is about 60.
What if we roll up all the way to 1-digit zip codes? Now things are more evenly distributed. The smallest region is 5 with a population of about 17.5M and the largest is 9 with a population of about 54M. A ratio of about 3.
Hey John,
I’ve never seen this truncation “in the wild”. Is it hidden inside APIs?
Yes, I was working with such a truncation yesterday. But it’s uncommon.