Chicago Scores 4.6% of Its Civic Notices. It Scores 54.2% of Its Agendas — and Not on the Same Fields.

The Setup

Chicago is one of only three metros — alongside San Francisco and Houston — where civic_records has ever received any of the three composite scores Locus attaches to civic documents: hostility_index, litigation_risk_score, and upzoning_probability. Every other tracked metro, Boston and Nashville included, sits at zero on all three. But Chicago's own coverage isn't spread evenly across its 211 records. 152 are notice filings (72% of volume); 59 are agenda filings (28%). Only 7 of the 152 notices, 4.6%, carry any composite score. 32 of the 59 agendas, 54.2%, carry upzoning_probability.

The Chain

The gap doesn't split cleanly by record type either — it splits within agenda between fields. All 32 scored agendas carry upzoning_probability, but only 18 carry hostility_index and litigation_risk_score together — 14 agenda records have an upzoning read with no matching hostility or litigation figure. Notices don't show that split: the 7 scored notices carry all three fields in lockstep, never one or two alone. Two different scoring behaviors are running against the same table — one populating all three fields together on a small notice subset, another populating upzoning_probability on nearly double the agenda rows it populates the other two on.

The Implication

Notices are Chicago's formal public-disclosure filings — the record type built to surface exactly the community friction hostility_index and litigation_risk_score exist to quantify. They're also the record type the scoring pipeline reaches least: 145 of 152 notices, 95.4%, have never been scored on any of the three fields. The 125 H3 cells covering the Chicago metro (average composite score 62.9) draw whatever civic signal exists from the 39 of 211 records that happened to get scored — a set skewed toward the minority record type, and within that set, toward one field over the other two.

What to Watch

Whether the 14 agenda records carrying upzoning_probability without hostility_index/litigation_risk_score eventually get those two fields backfilled, or whether that gap is evidence of two disconnected scoring passes rather than one incomplete one. Whether Chicago's 145 unscored notices move toward a higher coverage rate as new notices arrive, or whether the backlog just keeps growing unscored.

Limitations

Snapshot at query time over the full record history, not a rolling window, so coverage rates could shift as older or newer records are (re)scored. This does not evaluate whether the scored and unscored records differ in content in a way that would justify the gap, only that they differ in count by record type and field. The tie to cell_scores is presented as context on downstream stakes, not a measured weighting pathway.

--- Data as of 2026-09-11. Source: Locus `civic_records`, metro_slug='chicago' (211 records: 152 notice, 59 agenda; hostility_index/litigation_risk_score/upzoning_probability coverage by record_type), `cell_scores` metro_slug='chicago' (125 cells, avg composite 62.9).