Press kit — Ontario municipal elections 2026
The numbers a newsroom needs at deadline, with denominators and receipts. As of August 10–11, 2026; nominations close August 21, 2 p.m.; everything re-stamps at registrar-final. Companion pages: State of the Election · the Platform Baseline · the exceptional platforms · every race.
Headline figures
| figure | value | denominator |
|---|---|---|
| races total | 2,289 | — |
| zero-candidate races | 316 | 2,289 |
| acclamation-track races | 956 | 2,289 |
| contested races | 917 | 2,289 |
| candidates on file | 4,751 | — |
| researched/dispositioned | 4,743 | 4,751 |
| with campaign website | 1,267 | 4,743 |
| SCORED platforms (v2.1, full-field completion) | 416 | 4,743 |
| field with nothing readable (NPP+BROCHURE) | 4,074 (85.9%) | 4,743 |
| exceptional page names | 74 | 416 scored |
The methodology, in one paragraph
We assembled Ontario's full 2026 municipal candidate registry (4,751 candidates, 2,289 races, every directly-electing municipality), individually researched 4,743 of them to a dated conclusion (website found, or a recorded search that found none), captured and indexed every reachable campaign site, and read every candidate's captured content individually under a public, versioned, viewpoint-neutral instrument (METHODOLOGY v2.1): a candidate is only counted as having a platform if a reader can extract at least one verbatim forward commitment — a pledged future action with a number, mechanism, or timeline. 20.2% of the primary reads were blind double-rated by independent raters (90.8% class agreement post-adjudication, all disagreements adjudicated on the record); every rate we publish carries its denominator on the same line. No position is ever scored for its direction: a fully specified plan to build a thing and a fully specified plan to remove it count identically.
Downloadable tables
- table1_field_classes.csv — What a voter can read, class by class (n=4,743 researched)
- table2_races_registry.csv — Every race: municipality, office, seats, registrants (registry facts)
- table3_headline_figures.csv — Headline numbers with denominators and receipt citations
- table4_web_presence_by_office.csv — Website/platform coverage by office sought
Issue slices — every candidate touching each issue domain (14 CSVs)
- issue_arts-recreation.csv
- issue_climate-environment.csv
- issue_economy-jobs.csv
- issue_equity-inclusion.csv
- issue_governance-transparency.csv
- issue_growth-planning.csv
- issue_health-social.csv
- issue_housing.csv
- issue_roads-infrastructure.csv
- issue_rural-agriculture.csv
- issue_safety-policing.csv
- issue_schools-education.csv
- issue_taxes-spending.csv
- issue_transit.csv
Limitations, written by us first
Web-only record — print and door-to-door literature is invisible to this method, and much of small-town Ontario campaigns that way; absence of a published platform is a fact about the public record, not about a candidate's effort or worth. Snapshot dated 2026-08-10/11, ten days before nominations close (Aug 21, 2 p.m.) — the field will grow and we re-measure after close. 49 candidates' sites were unreachable to every fetch method we have, with per-site receipts; they are counted as their own class, never as "no platform." Scoring uses AI raters under an anchored public rubric with published reliability statistics and human lead adjudication — disclosed, reproducible, not eliminated. Per-candidate scores are never published; aggregates only, praise-only naming, and no aggregate is attached to a class of two or fewer identifiable people. Candidates named on the exceptional page passed three gates including an independent second reading; every one is quoted verbatim from their own platform with a link to it.