SONAR

Race standing from geo-confirmed signal

SONAR measures what residents of a place are actually saying about a race, then locks and scores every prediction in public.

It is not a poll. It is not a scrape of raw volume. It is a measured, weighted read of confirmed local accounts, published against a record that shows every miss next to every hit.

Snapshot as of loading. Static figures, not a live feed.

-
Markets deployed
-
Confirmed accounts
-
Platforms
-
Enriched signals

The mechanism, in three steps

01

Confirm residents

An account joins the measured pool only on evidence it belongs to the place: stated residence in the profile, membership in a local community, or a comparable resident signal. A geo-tagged post is not confirmation. Diaspora accounts are rejected.

02

Weigh impact

Each confirmed account contributes by realized reach and engagement, not by post count. No single account may exceed five percent of a pool's weight. Thin or concentrated pools are reported as thin or concentrated, not smoothed over.

03

Publish the record

Before an election, the standing is locked with a content hash and a timestamp. After the result certifies, the lock is scored against it. Both the prediction and the error are published. Misses appear at the same volume as hits.

What SONAR does not do

No demographic modeling. No turnout projection. No standing verdict in any market without a calibration record. Where the pool cannot support a call, the standing reads inconclusive with its reasons stated, and nothing louder.

The honest version of a number is worth more than a confident one. See the methodology for the full spine, and the calibration record for the ledger.