We need to run the synthetic PUFs through the taxdata routines to produce an equivalent file to the taxcalc puf.csv file, i.e. drop aggregate records, add CPS nonfilers, etc.
Doing this will allow us to run the synthetic files through taxcalc and compare it to the cps.csv file.
Based on the taxdata README, it looks like this could be done by cloning the repo, copying the synthesis to puf_data/StatMatch/Matching/puf2011.csv (per .gitignore), and then running make puf-files.
@andersonfrailey is this right? Or would you be able to run this for the two main candidate synthetic pufs?
We need to run the synthetic PUFs through the taxdata routines to produce an equivalent file to the taxcalc
puf.csv
file, i.e. drop aggregate records, add CPS nonfilers, etc.Doing this will allow us to run the synthetic files through taxcalc and compare it to the
cps.csv
file.Based on the taxdata README, it looks like this could be done by cloning the repo, copying the synthesis to
puf_data/StatMatch/Matching/puf2011.csv
(per.gitignore
), and then runningmake puf-files
.@andersonfrailey is this right? Or would you be able to run this for the two main candidate synthetic pufs?