org name matching

[mcneillucy]

Here is a list of things that are still incorrect, things that should continue to be worked on, and things to keep in mind within my rules! In the orgnames R script there is a #Testing section that should give good tools that I used to create my rules.

[mcneillucy]

ORG.COMMENT *filtered by rulemaking

• completed:
  FWS, NPS, FDA, EERE, EPA, VA, IRS, CFPB, OSHA, ATF, OFCCP, ACF, ETA, DOD, EEOC, PHMSA, CDC, BSEE, EBSA, OCC, DOI, MMS, EBSA, CPSC, LMSO, OPM, OSM, CMS, ED ,DOD, ETA, BLM, FNS, USCIS (0), FAA (0), NOAA

• Some agencies were too big to look through using the matching dockets within all comments....
  BLM, WHD, HHS

• some agencies were pretty messy
  NHTSA, NRC

• had problems loading as groups, but I assume these are fine on their own but couldn't get to them
  BOEM, DOL, OTS, HHS, NRC, BLM, DOD, CMS, ETA, FNS

• Notes:

  • played it safe after I completed general org.comment coding. There is lot of coding that just pertains to one agency that could have potential to be broadened, ex. lots of times when I str_dct(org, ".*") I do it for a bunch of agencies after I glance through them
  • there was difficulty deciphering within mass comments if there was one that was directly from the org, so there will be some missing TRUES there
  • knew less when I did the first few agencies so there may be some easy code to capture more, just can't tell with loading speed in the last few days, would definitely at least check back to make sure nothing is miscoding badly (ie. EPA, NPS, FWS)

[mcneillucy]

POSITION (two per docket)

• completed:
  EERE, NPS, FWS, FDA, VA, OSHA, ATF
  (note: we decided to wait for spreadsheet for EPA)

[mcneillucy]

ORG

• for the most part has been working well, with a few gathering more words before them than they should, but should be easy to remove (ex. attended the art institute)

• created script for one word orgs that are getting dropped because of a coding choice (line 255), these need to be added and/or use code (line 1069) after to bring the organization variable back into org if the org.comment == "T"

• Problems:

  • (line 250) org names with lots of "." are dropping out because of code that gets rid of names like J. Smith that are gibberish in org...losing names like N.R.A, A.O.I.B.P, and I.A.M National Pension Fund
  • there is large mismatching in org with firstname lastname in org
  • there are sometimes multiple orgs commenting together and that may not be captured in org