🧠 Knowledge Transfer

Current problem:

the review table is outputted then ingested again
super custom tables for Lynn

the point geometry for colp is bbl, but we do a lot of qaqc on address. two data products in one repo qaqc: the qaqc process is at the end (export_qaqc.sql, we do have description of the tables). but we also have a lot of flags created for this table.
Input data get geocoded twice
- ipis_colp_georesults
- geocoding + ipis -> so that we can qaqc

maybe model after facDB?
- currently we have two geocoding script
- one for qaqc (addresses)
- one for colp (bbl -> translated based on if it's air-rights or not)
- we should follow the model of facdb to store both geocoding in the same table
- then use sql to pick an choose
- easier for qaqc
- the only tricky part is that it's not easy to do in python because of air-right lots
- it's not source data to geocoding
qaqc for the address is for the input data
- qaqc on colp
- qaqc on ipis
- correcting colp
- or correction ipis
  
  we are producing colp, then it opened up an opportunity for a bunch of qaqc table (anything we can do with this data, so it's messy)
GOAL with QAQC #173
- qaqc to improve colp
- qaqc to improve ipis
- qaqc to improve geosupport
‼️ Changes to COLP leads to changes to facdb
Consolidating geocoding

combine two geocoding tables into one, so we only geocode once

Make string replacement logic more clear
turn it into a procedure
or maybe do it in python?