Shame Index — Grading Methodology

Version 0.1 · Draft · 2026-08-29 · Next review: pre-launch of live grades

1. What the grade measures

A Fault Line grade summarizes how a municipality performs on fixing reported infrastructure problems within its own stated timelines and against peer benchmarks. It is a performance measure of a government function, not a judgment of the community or its residents.

Grades apply to the municipal government (e.g., Boston Public Works) — not to elected individuals — and cover the categories Fault Line accepts reports for (potholes, streetlights, sidewalks, signage, drainage, and 19 other infrastructure categories).

2. The four inputs

The letter grade is a weighted composite of four measured inputs, each computed over a rolling 90-day window:

Each input is normalized against the peer group (see §4), scored 0–100, and combined by weight. The composite maps to a letter using the standard cutoffs: A ≥ 90, B ≥ 80, C ≥ 70, D ≥ 60, F < 60. Plus/minus modifiers apply at the top and bottom third of each band.

3. Minimum-data thresholds

A city is not graded until all four thresholds are met:

• Volume: ≥ 100 verified reports in the 90-day window.
• Coverage: Reports from ≥ 10 distinct reporters (protects against a single high-volume user distorting the sample).
• Category breadth: Reports across ≥ 3 distinct infrastructure categories.
• Escalation opportunity: ≥ 30 reports that have reached community-verification threshold and been escalated to the authority.

Cities that do not meet all four are shown as "Insufficient data" rather than assigned a grade. We consider a below-threshold grade less useful than no grade.

4. Peer groups

Cities are benchmarked against peers of comparable population and density, using U.S. Census-defined bands (small, mid-size, large urban; suburban; rural). A city's raw performance number is meaningful only against a peer group — a 48-hour median response is excellent for a small town and average for a major metro.

5. Data sources

• Fault Line report database: primary source for volume, severity, escalation, and recurrence.
• Authority responses: ingested from Open311, SeeClickFix, direct authority email replies, and, where available, municipal work-order APIs.
• Repeat-photo verification: user-submitted before/after photos at the same GPS confirm or reject a "fixed" state.
• Census TIGER boundaries: used to attribute each report to the correct municipal jurisdiction.

6. Update cadence

• Underlying metrics: recomputed nightly.
• Published grades: refreshed weekly. This dampens noise from single-day spikes and gives authorities a stable weekly window to respond to open issues.
• Methodology revisions: versioned. This page always shows the current version; prior versions are preserved and linkable so a grade issued under an older methodology can be reproduced.

7. What we deliberately do not do

• We do not grade elected officials individually. The grade measures the executive/administrative function, not an individual's voting record.
• We do not use social-media volume, news sentiment, or "vibes" inputs. Only measured performance.
• We do not assign grades below the data threshold to appear more comprehensive than we are.
• We do not accept municipal payments to change, delay, or suppress a grade. Cities can respond, dispute, and provide data — but not buy their grade.

8. Corrections and disputes

Municipal officials, authorities, and reporters can dispute inputs. A dispute channel accepts:

• Evidence a report was miscategorized or duplicated.
• Evidence a "fixed" state was missed (e.g., the work order was closed but not reflected in our system).
• Evidence a jurisdiction was mis-assigned (state road vs. city road is a common one).

Corrections are applied within the next weekly recompute. All accepted corrections are logged publicly on the affected city's grade page.

9. Limitations we're transparent about

• Reporter geography bias: neighborhoods with more Fault Line users generate more reports. We publish reporter-density heatmaps alongside grades so the reader can weight for this.
• Jurisdictional handoffs: a pothole on a state highway inside city limits is often not the city's to fix. We route to the correct authority and do not penalize the city for delays outside its control — but attribution errors happen and are the leading source of disputes.
• Beta baseline: until a peer group has enough graded cities to normalize against, early grades will use national defaults and be marked as such.

10. Not legal or investment advice

Grades are provided for civic transparency and public discussion. They are not a rating in the sense used by financial-services regulators and should not be used as the sole basis for procurement, investment, or legal decisions. Journalists, researchers, and residents are welcome to cite them; we ask that citations link to this methodology page and note the grade version.

Contact

Corrections, disputes, or methodology feedback:
Email: email Moonlit Social Labs
Michael Wylde · Moonlit Social Labs