States without 2016 Presidential vote

[kuriwaki]

Only two states are missing 2016 Presidential vote:

• West Virginia
• Mississippi.

This is not really a bug because VEST, MGGG, openelections don't seem to have that data either. But just flagging here in case someone finds a solution. It would be great to have all 50-states.

It is worth noting the NYT Upshot has 2020 precinct results in GeoJSON for those WV and MS. They don't seem to have 2016 for those states, though.

[rootsmusic]

Besides those you've linked, another repo for election results is MEDSL. For 2016 Presidential results, OpenElections seems to have both MS and WV.

[kuriwaki]

Thanks. MEDSL and openelections don't have shapefiles; this repo's process matches by geometry. So we will need a precinct shapefile with precinct names that are similar enough to the MEDSL/openelection csv's.

[kuriwaki]

Re:

VEST's 2016 results were mapped on shapefiles. Hopefully I'm understanding this issue. (@rootsmusic)

That VEST link does not have a wv_2016.zip or a ms_2016.zip, which is the root of this issue.

So I'd like to procure those two states from other means. Like VEST, the product should have a shapefile and 2016 presidential election results on it.

• MEDSL/open-elections does have election results for those states, for the most part, but not shapefiles.
• The two states' websites have shapefiles (at least 2010 ones) but no election results.

To merge the two, we probably need to merge on name of precinct.

[kuriwaki]

I looked in a bit more into the data.

Mississippi

• shapefile: ✅ A state agency has a 2010 and 2020 VTD shp file: https://www.maris.state.ms.us/HTML/DATA/Political.html#gsc.tab=0
• MEDSL: ✅ seems to have right number of precincts
• Precinct naming structure: County + name (e.g. county FIPS 28017, Plesasant Grove)
• Merge: left_join on county and precinct links 57% of the 1800 precincts. The rest maybe could be done with careful inspection of the name formats.

West Virginia

• shapefile : ❓ I can only find a 2010 VTD shpfile: http://www.wvgis.wvu.edu/data/dataset.php?ID=433 boundaries may or may not have changed since 2010.
• MESDL: ❓ In a few counties they are missing precinct-level results in at least 400 out of ~1800 precincts, and instead report only the county total. Those counties are Kanawha, Cabell (two of the largest counties in the state), Nicholas, Wyoming, Marshall, Webster, Doddridge counties. Fortunately, Openelections WV data has csvs of precinct-level election results for all but one those counties. We could stack.
• Precinct naming structure: they use numbers (Kanawha county, 34A), not name
• Merge: Currently, merge on county and precinct number links 66% of the 1800 precincts.

I can try to submit a PR for each state?

[rootsmusic]

Precinct shapefiles aren't readily available from many states, and I strongly advise validating completeness even if available. The Census has 2020 VTD shapefiles (for P.L. 94-171), but VTDs may not be equivalent to precincts. If they aren't, then hopefully the state (which may or may not be from its board of elections) can provide a crosswalk.

P.S. if you open an issue in MEDSL's repo about their incomplete results, their staff will work on it.

[kuriwaki]

Ah yes, Cory and Chris already use the PL94171 as the basis here (via their PL94171 package), and I see that the VTD names in there are similar to what I found each SOS dataset. So I'll just use the package.

Not sure what "validating" with precinct files would entail though, because I don't have a VEST precinct shapefile. We can certainly check the vote totals add up and we can do a visual inspection with Upshot's "Extremely detailed map of 2016".

[rootsmusic]

Yes, basic validation is just whether the shapefile is complete. I advise validating precinct shapefiles because VEST doesn't have time to validate boundaries. OpenPrecincts isn't active for 2020 elections, but they have a Python script for verifying shapefiles.

[christopherkenny]

Yes, we've already updated these along with the elections from KY.