District Data
Intake Checklist

Cross-reference the TIGER district shapefile against county boundaries. Districts can span multiple counties. Some townships split across district lines.

Format: .txt (pipe or comma delimited). Same 136-column schema for all counties. File named by county FIPS. Request ACTIVE + INACTIVE + CONFIRMATION statuses.

Already in database: CLERMONT BROWN

Append to existing voters table. Run scoring model on new records. Add county to index. Verify row count matches SOS file.

Preferred format: CSV or Excel export from county CAMA system. Most Ohio counties use Tyler Technologies or Aumentum — exports are similar. If no bulk download: call auditor's office and request public records extract.

Clermont: maps.clermontcountyohio.gov  |  Brown: received via auditor export

Minimum required fields. If sale data is missing, it may be in a separate transfer file. Land use code distinguishes residential dwellings from vacant land — critical for unregistered household count accuracy.

Critical step. Brown County had 7.4% match rate due to inconsistent address formats (missing street types). Target 50%+ match rate. If below 30%, request standardized export or use OGRIP API as supplement.

Use the OGRIP fetcher tool to pull standardized addresses if county data is poor.

Append to parcels table. Run voter-parcel match. Residential parcels with no matched voter → unregistered_households table. Verify counts — residential parcel count should not far exceed voter count.

Usually a .shp or zipped shapefile set. Clermont: ClermontAddresses.zip from county GIS. Brown: Addresses_ft.zip from engineer's GIS site. Check coordinate system — may need reprojection to EPSG:4269.

Clermont had pre-computed X (lon) and Y (lat) fields in the DBF — use those directly. Brown required reprojection from Ohio State Plane Feet (EPSG:3735) to WGS84 (EPSG:4269) via geopandas. Check CRS in .prj file.

Uses exact match + street interpolation (house number between known points). Below 30% means address formats between parcel file and address points file are too different — escalate to OGRIP or call county auditor.

Clermont: VotingPrecincts.shp from county GIS. Brown: VOTING-PRECINCT.zip from engineer GIS. Other counties: check county BOE site or GIS downloads page. Field name for precinct name varies — check DBF fields after loading.

Reproject precinct shapefile to EPSG:4269 before joining. Match rate limited by geocoding rate. Records without coordinates get INFERRED - VERIFY flag. Use assign_precincts.py script (already built).

Mismatch is common — shapefile may say BATAVIA TWP A while voter file says BATAVIA TOWNSHIP A. Build a crosswalk if needed. Villages are separate precincts in voter file — confirm they appear in shapefile too.

Ohio SOS publishes county voter registration totals monthly. Cross-check your database count against their published number. Discrepancy over 1% warrants investigation.

Residential parcel count should not far exceed registered voter addresses. Rule of thumb: unregistered count should be less than total registered count. If unregistered > registered, address matching is failing — investigate before using for canvassing.

Pick 10 known addresses across the district. Confirm their precinct and district in both the database and the county BOE website. This catches CRS mismatches and boundary errors that summary stats miss.

Never deliver a report without disclosing what percentage of precinct/district assignments were confirmed by spatial join vs. inferred. Candidate teams need to know which addresses to treat as verified vs. estimated.

One Excel or CSV file per district. Includes: total voters by precinct, Soft R + Independent scores, dropout count, new mover count, unregistered count. Named HD[NN]_Deliverable.xlsx.

Separate CSV per district for field operations. Sorted by precinct then address for walk list efficiency. Named HD[NN]_TargetVoters.csv. Confirm with candidate before distribution.

Sorted by sale date descending — newest arrivals first. Include sale price and assessed value. Named HD[NN]_WelcomeWagon.csv. This list refreshes monthly as new sales are recorded.