Add ACS, rent and property taxes and 3-year CPS

[PavelMakarchuk]

Fixes #31

[MaxGhenis]

Have you tried running it?

[MaxGhenis]

Could you add acs/README.md linking to the codebook?

Also if easy, could you add rent and property tax as variables?

[PavelMakarchuk]

Could you add acs/README.md linking to the codebook?

Also if easy, could you add rent and property tax as variables?

Added a readme with documentation

We already include
GRNTP -> Gross Rent
RNTP -> Monthly rent

I don't see variables for property taxes in the dictionary

[MaxGhenis]

[PavelMakarchuk]

[broken image removed by @nikhilwoodruff- weird gh issue?]
No idea how i missed that - added

[nikhilwoodruff]

@MaxGhenis a few things (thanks @PavelMakarchuk; taking over the baton for this PR) you might want to take a look at (but not essential or blocking).

• I've moved from our implementation of QRFs in survey-enhance to a more lightweight wrapper over quantile-forest which seems faster and better-written than mine. We still have the mean quantile ability. The new implementation wrapper is in utils/qrf.py.
• Seems like we shouldn't calibrate real estate taxes from the SOI? But I can add calibration targets for total rent and property taxes, any ideas on the definitive source to use? They match the ACS totals after imputation.

[MaxGhenis]

Correct we should not calibrate real estate taxes from SOI. See #29 for other options.

[nikhilwoodruff]

SNAP is looking really nice with the rent imputation: after applying USDA's 82% take-up rate, we get 103bn vs 107bn in 2023.

[nikhilwoodruff]

Also tied in #67 here.

[nikhilwoodruff]

Getting memory breaches but I'm going to merge since it ran locally.