How we score suburbs

Every weight, every sub-weight, every scale, every grade band and every source, on one page, linked from every place page. The weights are an editorial judgement and we would rather be argued with than trusted.

The seven categories and what goes into them

Each input is converted to a 0–100 score on a fixed, published scale, so adding a new metro never moves an existing place. Two kinds of input are relative on purpose: home value, rent and income are compared with the metro they sit in, and school proficiency is a percentile among districts in the same state. A category needs at least half its input weight to be scored. The overall score is the weighted mean of the categories that have a score, re-normalised to 100%.

Housing & affordability
18%
Median home value vs metro50 at the metro figure; higher is better; 0/100 at 2.5× (log)40%
Median rent vs metro50 at the metro figure; higher is better; 0/100 at 2× (log)30%
Ownership rate0 at 35% → 100 at 90%15%
Households spending >30% of income on housing0 at 45% → 100 at 15%15%
Schools
18%
Within-state proficiency percentilepercentile used as the score50%
Student–teacher ratio0 at 24.0 → 100 at 12.020%
Graduation rate0 at 75% → 100 at 100%20%
Schools within 3 mileslog scale, 100 at 30+10%
Cost of living
14%
Housing cost index0 at 200 → 100 at 7045%
Effective property tax rate0 at 2.40% → 100 at 0.40%25%
Utilities0 at 160 → 100 at 6020%
Local sales tax0 at 10.00% → 100 at 5.00%10%
Commute & transport
14%
Mean commute time0 at 40.0 min → 100 at 18.0 min45%
Share commuting under 30 min0 at 30% → 100 at 80%25%
Transit stop densitylog scale, 100 at 25+10%
Share working from home0 at 3.0% → 100 at 30%20%
Amenities
12%
Grocerieslog scale, 100 at 25+25%
Restaurants & caféslog scale, 100 at 250+20%
Parkslog scale, 100 at 60+20%
Fitnesslog scale, 100 at 30+15%
Healthcarelog scale, 100 at 60+10%
Retaillog scale, 100 at 200+10%
Employment & income
12%
Median household income vs metro50 at the metro figure; higher is better; 0/100 at 2.5× (log)40%
Unemployment rate0 at 9.0% → 100 at 2.0%30%
Labour force participation0 at 50% → 100 at 78%20%
Share in professional occupations0 at 20% → 100 at 70%10%
Residents
12%
Twelve poll questions, equal weight — mean share answering yes100%
Responses needed before the category counts8

Thirteen grade bands, five points each

A score is only useful if the letter attached to it is defined. F absorbs everything below 38 because pretending to resolve it would be false precision.

F
0–37
D−
38–42
D
43–47
D+
48–52
C−
53–57
C
58–62
C+
63–67
B−
68–72
B
73–77
B+
78–82
A−
83–87
A
88–92
A+
93–100

Sources, and when each was last true

Census Gazetteer — places2025
Place GEOIDs, land area and internal points, used for every radius and distance on the site
Public domain (US federal government) · source
Census American Community Survey, 5-year estimatesACS 2020–24
Income, home value, rent, owner costs, tenure, housing cost burden, real estate taxes (aggregate taxes ÷ aggregate value gives the effective property tax rate), commute time and mode, employment and occupations, and — as unscored context — the median year built, the share of houses rather than flats and the share of households with children, at Census place level and for the metro area
Public domain (US federal government) · source
NCES Common Core of Data + EDGE school locationsSY2024–25
Every open, regular public school with its district (LEAID), enrolment and teacher FTE. Used for schools within 3 miles and the student–teacher ratio
Public domain (US federal government) · source
EDFacts state assessments and adjusted cohort graduation ratesSY2021–22
District proficiency in mathematics and reading (SY2021–22), converted to within-state percentiles and never compared across states. Four-year adjusted cohort graduation rates (SY2018–19)
Public domain (US federal government) · source
OpenStreetMapSep 2026
Groceries, restaurants and cafés, parks, fitness, healthcare, retail and transit stops within 2 miles, held in a separate layer
© OpenStreetMap contributors, ODbL 1.0 · source
Sales tax rates2026-09
Combined state and local sales tax. Place-specific where the state publishes a city rate file (Texas Comptroller, California CDTFA, Washington DOR); otherwise the state rate plus the population-weighted average local rate from the Tax Foundation
Public record / Tax Foundation · source
Cost-of-living index (composed in-house)Sep 2026
No free national index exists at this granularity. Composed from housing cost (owner costs and rent against the metro, blended by tenure), utilities (gross minus contract rent), effective property tax and local sales tax
Suburb Grade
Resident pollLive
This site's own poll. Twelve yes/no questions, equal weight, eight responses to activate
Suburb Grade
The launch-day rule

A suburb with fewer than 8 poll responses has no Residents grade. The category is excluded and the other six re-normalise to 100%, and the place page says so in words rather than leaving a blank cell. 1,795 of the 1,795 places are currently below the threshold. Once a place activates, its headline number depends partly on the people who live there — the part of this index nobody can copy from the same public sources. Activation is dated and logged on the place page.

The twelve questions
  1. Is it quiet at night?
  2. Do you feel safe walking after dark?
  3. Can you walk to a grocery store?
  4. Can you walk to a restaurant or café?
  5. Is it good for young children?
  6. Is it good for teenagers?
  7. Are the neighbours friendly?
  8. Are the sidewalks in good condition?
  9. Are yards and streets well kept?
  10. Is parking easy?
  11. Is the internet reliable?
  12. Would you move here again?

One tap each, no account. A respondent is an anonymous first-party cookie; changing an answer replaces it rather than adding a vote, and new respondents from one network are capped per place per day.

What we deliberately don't score
—Crime. FBI data floors at the police agency, so every neighbourhood in a city would carry the city's figure. Poll question 2 carries the safety intent instead.
—School-level pages. Four established sites hold that result page completely. No entry point worth the build.
—Cross-state comparison. States set their own tests and cut scores. A national proficiency score is confidently wrong, so school proficiency is only ever a within-state percentile.
—Which school serves this address. The national attendance-boundary dataset is from 2015–16 and only about half of it is real catchments. We show schools near a place and name the district.
—Accounts and property listings. Neither tests the thing this pilot exists to test.

How the written pages are produced

The scores are arithmetic. The paragraphs next to them are writing, and you should know how they were made.

Drafted by a language model, checked mechanically, published by a person
— Each page is drafted by a language model from a brief containing nothing but the figures on this site, with their sources. The model is given no other knowledge of the place, and is told it may not add any.
— Every draft is then checked by a program against that brief. A number that is not in the brief and cannot be worked out from it fails. So does a place name, road or district that the brief never mentions — an invented neighbourhood is the most plausible-sounding mistake a model makes, and the check exists because one got through in testing.
— Nothing is published automatically. A page stays a draft, and stays hidden from search engines, until a person has read it. Pages carry the model that drafted them in their source.
— What this cannot do is tell you what a street feels like. Where the data is silent — the age of a particular house, the school zone for one address, what the commute is really like at 8am — the pages are meant to say so rather than guess. If you find one that guesses, it is a bug.

List weightings

Each ranked list has its own weighting, published on the list and repeated here. None of them is the overall score.

Best for families
Schools 30%Amenities 20%Resident poll (Q2, Q5, Q6, Q8) 20%Cost of living 15%Housing & affordability 10%Commute & transport 5%
Most affordable
Cost of living 55%Median home value (lower is better) 25%Cost-burdened households 20%
Best for schools
Schools 80%Resident poll (Q5, Q6) 20%
Shortest commutes downtown
Distance to downtown 50%Mean commute time 50%
Most walkable
Amenities 40%Transit stop density 15%Resident poll (Q3, Q4, Q8) 45%
Best for young professionals
Employment & income 25%Amenities 25%Commute & transport 20%Median rent (lower is better) 15%Cost of living 15%