Closed sastoudt closed 5 years ago
I think this version is now technically correct. It is missing a PUMS example, but I think there needs to be a separate function for that. The current approximation would work, but you could do better by just using the individual data you have from PUMS instead of inputting bins. If we merge this PR, then I can start fresh and break things up into approximate_median_acs
(wouldn't need the design factor anymore) and approximate_median_pums
.
Based on census feedback: