Voter Targeting Universe
Aggregate-only district scoring for Michigan races. Precinct and segment rollups of turnout propensity, partisan lean, and universe assignments (GOTV-HIGH, GOTV-MED, PERSUASION, BASE, OPPOSE, LOW-PRIORITY). Built on the Michigan Qualified Voter File and the turnout + partisan models I run. No individual-level records leave the shop.
Universe assignment
Each voter in the district is scored on turnout propensity and partisan support, then assigned to one of six universes. Deliverables aggregate the assignments to precinct, district, and demographic segment levels. No voter-level records are exported. Thresholds are configurable per-race in the engagement.
| Universe | Logic | Action |
|---|---|---|
| GOTV-HIGH | High D support + low turnout | Mobilize: highest contact priority |
| GOTV-MED | Moderate D support + low turnout | Mobilize with persuasion messaging |
| PERSUASION | High turnout + moderate support | Persuade: they will vote, need convincing |
| BASE | High D support + high turnout | Low-touch: reliable supporters |
| OPPOSE | Low D support | Do not contact |
| LOW-PRIORITY | Very low turnout | Not worth resources |
Score definitions
- Turnout propensity: P(votes in the named primary). XGBoost classifier; AUC ~0.98 on held-out cross-validation. Top features: primary-participation ratio, permanent-AV status, total-elections count.
- Partisan lean: D/R lean using ecological inference from precinct results plus individual behavioral signals. Michigan does not have party registration; lean is estimated, not declared.
- Support score: campaign-specific support estimate. Defaults to partisan lean (no campaign engagement); upgrades to endorsement-proxy (FEC/MiTN donor overlap) or candidate-specific (trained on canvass results) when available.
- GOTV priority: composite score combining turnout propensity and support to rank GOTV contact order.
- Persuasion priority: composite score for persuasion targeting.
Deliverables
ZIP package containing aggregated targeting outputs, README, methodology, and a signed manifest. No voter-level records.
- universe-by-precinct.csv: count of voters in each universe per precinct, with turnout history columns
- district-summary.csv: district-level universe distribution, demographic breakdown, and turnout trend
- top-priority-precincts.csv: ranked precinct list for GOTV and persuasion deployment
- segment-summaries.csv: universe distribution by demographic segment (age, sex, race-proxy)
- maps: precinct-level choropleths of universe density (PNG and SVG)
- README.txt: engagement-specific notes (turnout target, model version, universe thresholds)
- methodology.md: universe definitions, score derivation, data sources
- manifest.json: cryptographic hash + delivery date for audit
Example output
A public-safe sample is available below. The sample uses real Michigan precinct geography and aggregate model outputs. No voter-level records are present in the deliverable or the sample by design.
Download sample ZIPPricing
Price scales by office level of the named race or district.
| Office tier | Price |
|---|---|
| School board / township | $1,500 |
| City | $2,500 |
| County / judicial | $4,000 |
| Michigan House | $6,000 |
| Michigan Senate | $10,000 |
| Statewide / federal | $20,000 |
Delivery timeline
- Standard: 14 business days from complete order.
- Rush: 8 business days, +50% on base price.
- Delivered as a download link to the ZIP package.
How to order
Email info@regionpulse.com with: race or district identifier, delivery email, and rush (yes/no). I will reply with a fixed-price invoice.
info@regionpulse.com