How CityCites builds every score
The principles behind every score: public sources, comparability, missing data, confidence and versioning. Each score also has its own page.
Principles
- Public sources only. The Census Bureau, FEMA, the Federal Audit Clearinghouse and the Bureau of Labor Statistics.
- Comparable. Dollar figures are adjusted for inflation, and ranked scores compare a city with other US cities on a 0–100 scale where higher is better. Hazard exposure compares each hazard with the typical US city, so a risk that is ordinary everywhere doesn’t drown out one that sets a place apart.
- Never fabricated. Missing data is excluded and disclosed, never estimated or filled in.
- Traceable. Every score carries its sources, source dates and methodology version.
How data becomes a score
- Census surveys
- Government finances
- FEMA hazards & flood maps
- Federal audits
- BLS inflation
- 1Acquire
- 2Normalize
- 3Validate
- 4Score
- 5Review & release
- City pages
- Dashboards
- API & data feeds
| Step | What happens |
|---|---|
| 1. Acquire | Public datasets and city financial documents are collected on each source's release schedule. |
| 2. Normalize | Every figure is matched to the right city, year and source. |
| 3. Validate | Automated checks catch gaps, unexpected zeros, category changes, duplicates and unusual year-over-year moves. |
| 4. Score | Inputs are inflation-adjusted, compared nationally and combined into component and overall scores. |
| 5. Review & release | Each release is reviewed, versioned and published. Dashboards, city pages and the API all read the same release. |
Peer groups
Peer percentiles compare a city with cities of similar population: under 10k, 10k–50k, 50k–150k and 150k+.
Missing data & completeness
CityCites never fills a gap with a guess. If an input is missing, its component is left out and the score is calculated from what remains. Each score reports its data_completeness so users can see how much of it was available. A score with too little data is reported as unavailable.
Confidence
Every score is labeled high, medium or low confidence, based on data completeness and, for survey-based inputs, the size of the city.
Display policies
standard scores may be shown to consumers with a link to their definition. research data (the Demographic Diversity Index) is not displayed in consumer housing contexts by default; partners' compliance teams decide. The political lean indicator is descriptive: never ranked or part of a score, and delivered only to partners who agree not to use it in housing, lending, insurance, employment or advertising decisions.
Versioning
Methodology is versioned (2027.3). Any change to inputs, weights or scoring creates a new version, and every score reports which version produced it.
Technical methodology
Technical methodology for partners
Data partners receive the full technical methodology: exact inputs, time windows, weights, normalization and validation rules, so their model-risk, compliance and data-science teams can review every score.
Request the technical methodology →About this preview
What this preview includes
- 8,099 US cities scored from public data downloaded 2026-11, each with a public profile page.
- Community Investment's execution component is not available nationally yet; it is left out and the score shows reduced completeness.
- The API shown is a static sample for one city, not a live service.