How We Build
Every City Profile
The exact sources, formulas, update schedule and limitations behind the numbers on UrbFact. If it is on a city page, this is where it comes from.
From raw data to a 1 to 10 score
Each city gets six category scores (Economy, Housing, Safety, Education, Weather, Demographics) and one overall score. A score is not an opinion. For every metric we take the city value and rank it against the national distribution of all US cities, then convert that percentile to a 1 to 10 scale and an A+ to F letter grade.
In plain terms: a city in the 90th percentile for median income scores near the top on that metric, while one in the 20th percentile scores low. Category scores are the weighted average of the metrics inside them, and the overall score is the average of the six categories.
Metrics where lower is better (crime rate, poverty rate, unemployment, cost burden) are inverted before ranking, so that safer and more affordable always means a higher score. Cities missing the data for a metric are excluded from that metric rather than assigned a zero, so a gap never silently drags a score down.
Where each number comes from, and how current it is
Every figure traces to an official dataset. Below is the source, what it covers, and how often we refresh it.
US Census Bureau
American Community Survey 2023, 5-year estimates (per place)
Population, income, poverty, education, housing, commute and household data. Refreshed each year when the new 5-year release is published.
FBI
Crime Data Explorer, calendar year 2024, agency level
Violent and property crime per 100,000 residents from each city's own police agency, used only when the agency reported at least nine months of the year. Offense subcategories (murder, robbery, burglary and others) are shown for cities of 25,000 and more. Refreshed each year when the FBI publishes the prior year.
NOAA
Climate Normals 1991-2020, monthly
Monthly temperature, precipitation and snowfall from NCEI 1991-2020 normals, taken from the nearest weather station within 40 miles (airport stations preferred). Outside the US: WorldClim 2.1 normals (1970-2000) and NASA POWER.
EPA & CDC
AirData annual AQI 2024, PLACES 2024
Air quality index by county from EPA monitors (only counties with a monitor). Obesity, diabetes, smoking, insurance and other health measures per place from CDC PLACES. Unemployment comes from the ACS.
HUD & Census Housing
Fair Market Rents FY2026, ACS 2023 Housing, Zillow ZHVI monthly
Median home value, median rent, homeownership and cost burden. Refreshed alongside the yearly ACS and HUD releases.
NCES & Census Education
Education attainment tables
High school and college attainment and school district context, from the education tables of the same ACS vintage.
What the data can and cannot tell you
Small places carry larger margins of error. For a town of a few thousand people, survey-based figures like median income or rent are estimates with a real confidence interval, not exact counts. We show them because they are still useful, but treat single small-city numbers as directional rather than precise.
Crime data depends on local agencies reporting to the FBI. Where an agency did not report, or reported only part of the year, the crime fields are left empty. We never estimate, interpolate or borrow a neighboring city's figure.
Outside the United States, comparable per-city government data is far thinner. For international cities we publish population (Wikidata dated census and official estimates, GeoNames), climate, elevation and geography, and leave income, housing and crime blank rather than filling gaps with a national average pretending to be a city figure. Population figures without a year are GeoNames values whose reference year is unknown.
See a number that looks wrong?
Data corrections matter to us. Tell us the city and metric and we will trace it back to the source.