How the ranking works
Every one of the 103 metros is scored 0–100 on each factor by percentile against all the others — a 90 means it beats 90% of US metros on that factor. Your slider positions become weights, and the final score is the weighted average of the factors you left above zero. No black box, no editorial thumb on the scale, no AI: move a slider and the arithmetic changes, that is all.
Live data ingestion
Five factor families are refreshed automatically from primary sources rather than held as static estimates:
- Income, population, commute, attainment. Census ACS 5-year metro tables (B19013, B25077, B01003, S1501, S1701, S0801), pulled by CBSA and matched to our metro list on city tokens plus state.
- Home values and rents. Zillow ZHVI and ZORI metro research files, monthly. Five-year growth is computed from the same series, so the appreciation number and the price number always agree.
- Job market. BLS Local Area Unemployment Statistics — the current metro unemployment rate plus its year-over-year change, blended 70/30 into a job-market index.
- Safety. FBI Crime Data Explorer estimates, violent crime weighted 3× property crime, inverted. State-level where metro reporting is incomplete.
- Population diversity. Census ACS table B03002 gives the racial and ethnic composition of each metro. We publish the breakdown as plain information and derive a Simpson diversity index from it — how evenly the population is spread across White, Black, Hispanic, Asian and other groups. It is descriptive rather than evaluative, and you can weight it to zero.
- School outcomes. NCES EDFacts grade-8 reading and math proficiency, published per school district and averaged across each metro weighted by how many students sat the tests, joined to metros through the NCES Common Core of Data district directory. That measured result carries 65% of the school factor; ACS bachelor’s attainment and child poverty carry the remaining 35% as environment context. Metros with no published district results fall back to the ACS proxy alone.
- School sports & recruiting. The NCES IPEDS directory of four-year colleges and universities, joined to metros by CBSA. Each institution is weighted by its enrolment band, because scholarship capacity rises steeply with university size, and the metro score blends the total capacity in the metro (45%), the capacity per resident (30%) and the capacity across the whole state (25%), since recruiting routinely travels statewide.
Every refresh is logged with its vintage per source; anything a source does not cover keeps the compiled baseline value and is never presented as live.
The factors
- Temperate climate
- Mild winters and summers that stay livable. Source: NOAA 1991–2020 climate normals.
- School academics
- Measured K-12 test outcomes, attainment and child poverty. Source: NCES / EDFacts + state assessments.
- School sports & recruiting
- How much college scholarship attention reaches local high schools. Source: NCES IPEDS four-year institutions by metro and state.
- Safe neighbourhoods
- Violent and property crime, inverted. Source: FBI Crime Data Explorer (coverage varies).
- Home affordability
- Typical home value against local income. Source: Zillow ZHVI + Census ACS.
- Property growth
- Five-year home value appreciation. Source: Zillow ZHVI.
- Faith community
- Congregation and adherent density. Source: US Religion Census 2020 (county, decennial).
- College pipeline
- Degree attainment, university access, recruiting exposure. Source: NCES outcomes + proximity proxies.
- Job market
- Unemployment level and its recent trend. Source: BLS Local Area Unemployment Statistics.
- Income vs. cost
- Median household income adjusted for local prices. Source: Census ACS + regional price parities.
- Self-employment health
- Small-business formation and survival. Source: Census County Business Patterns + SBA.
- Short commutes
- Mean travel time to work. Source: Census ACS.
- Population diversity
- How evenly mixed the metro's racial and ethnic makeup is. Source: Census ACS B03002 (race and Hispanic origin).
- Healthcare access
- Hospital quality and provider density. Source: HRSA + CMS hospital ratings.
Where the data is weak
Three factors deserve a caveat, and we would rather say so than pretend:
- School academics. The well-known 1–10 school ratings are commercially licensed and expensive. We use public NCES EDFacts assessment results instead, which are real test outcomes rather than estimates — but they lag by a few years, states set their own proficiency bars so cross-state comparison is imperfect, and a metro average is never a substitute for checking an individual school.
- Sports recruiting. Nobody publishes a free dataset of which high schools college teams actually sign athletes from — 247Sports, On3 and Rivals license theirs commercially. Our sports score measures the presence and scale of college programs within recruiting range instead. It is a genuine measure of scholarship opportunity nearby, but it is not a count of signings, and a strong individual high school programme can sit in a metro that scores low.
- Faith community. The US Religion Census publishes adherent and congregation counts by county, once a decade. It is a density measure, not a quality or denomination-fit measure, and the 2020 vintage is the newest that exists.
- College pipeline. There is no licensable dataset of athletic recruiting intensity. This factor combines degree attainment, proximity and capacity of in-state universities, and college-going rates. It does not claim to measure how many D1 scouts visit a given high school.
Crime data also carries a caveat: FBI reporting is voluntary and some large agencies report incompletely, so safety scores blend federal data with local open-data portals where available.
Refresh cadence
Home values and rents refresh monthly from Zillow and Redfin research data. Income, commute and business counts refresh annually with Census ACS and BLS releases. Climate normals change once a decade. Religion Census data changes once a decade. Each figure carries the vintage of its source, not the date you loaded the page.