WIP refactor

[SPTKL]

Refactor

ACS - CDTA #105

• for acs, we are taking data from db-factfinder acs_community_profiles pipeline
  https://github.com/NYCPlanning/db-factfinder/blob/dev/pipelines/acs_community_profiles.py
• not sure if we are taking MOE for some variables also, see Add floodplain MOEs for ACS vars #98
  https://github.com/NYCPlanning/db-community-profiles/pull/98/files#diff-f1e2480849c4c0b23e939fea64264a28eff4e9cbee18aa686350187188d08154

✅ Crime - CDTA via spatial join

• we only switched out the spatial boundary from cd to cdta, everything else remains mostly the same

Facdb - CDTA via spatial join

• since the version of facdb with cdta2020 is not published yet, we are using spatial join to get cdta2020 for facdb

would need to reconfirm methodology

Floodplain - CDTA

• the acs part of the data is coming from db-factfinder, the acs_community_profiles pipeline
• the decennial part comes from the published decennial pl release data Add pipeline to incorporate PL data #107. we use block level population and join with lookup_geo to calculate floodplain population

would need to reconfirm methodology

✅ Park accss - CDTA

• similar method as floodplain, we take block level decennial population and join with lookup_geo to calculate park access

pluto - CDTA via spatial join

• the pluto attributes are using spatial join, since cdta is not a field in pluto. We are currently using lat/lon, which lead to some records not assigned to a cdta (because point in water & etc)

would need to reconfirm methodology
Question: are we using shoreline clipped or unclipped, and centroid or polygon to do spatial join?

Poverty - CD (not refactored)

• currently the data is coming from a static csv stored in the data folder.
• the data is on CD level

• • Need to get updated data. Came from Population or Planning Coordination. Can we get at CDTA level?

✅ Sanitation - CD (not refactored)

• the santiation data from open data is cd level.

Static Data

• cb_contact
• cd_decennial_pop -> need from population
• cd_poverty
• cd_puma -> is puma still relevant?
• cd_son (cd and neighborhood names)
• cd_title
• cd_to_block -> how is this relevant? this table is not used anywhere, consider remove
• cd_tooltips -> someone would need to review this
• crimes_2019.csv -> probably would need to be crime 2020

[SPTKL]

#111

[AmandaDoyle]

Going to be a conversation with decision - use CD where we've had CD and CDTA where we've had PUMA?

[AmandaDoyle]

ACS and Floodplain should be good with how it is.

[AmandaDoyle]

Need to get official population counts from Population