Processing

Editing shapefile precinct names here, and election result precinct names here. Run those two scripts individually and afterwards, merge the clean shapefile and clean election results using this script (note: these generate CSV files for the ease of viewing mis-matches, but eventually you will want to save .shp and .geojson files to preserve the geometry column).

NJ released official election district boundaries in last September 2020. You can find them here: https://njogis-newjersey.opendata.arcgis.com/datasets/election-districts-for-new-jersey

It is worthwhile to try matching these boundaries to election results to see how it compares to the census partnership file.

Mismatches to address

Atlantic (001)

Election results “port republic city vote” should probably be deleted because I doubt that is a unique precinct. Verify with county

Camden (007)

Election results have “Gloucester City ward 3” and “Gloucester city W3”. What is the difference?
Election results have “City of Gloucester” and “City of Camden” precincts. Probably not unique precincts. Verify with county

Mercer County (021)

“Princeton” precincts: Shapefile has princeton township and princeton borough , elections just have “princeton”

Other issues

There are precincts in the election results that I suspect are not unique precincts. I believe they are the counts of precinct groupings aggregated into one row. For example, in Warren county, we see the following precinct names:

041,washtwp 5 041,washtwp 4 041,washtwp 3 041,washtwp 2 041,washtwp 1 041,washtwp

In this example, the final precinct, “washtwp” appears to be close to the sum of the precincts named washtwp 1-5. If this is truly the case, the final precinct should be removed from the election results.

In the document “mismatch_july2020.xlsx”, I highlighted the precincts in red that I suspect should be deleted.

PrincetonUniversity / gerryspam