Methodology

Our Approach

Repit aggregates public and proprietary data to give homebuyers and investors a clear view of housing markets at the ZIP, city, metro, and state level. We don't predict. We measure outcomes.

Every metric on this site is calculated from verifiable data sources — federal, commercial, and internal. Where we make assumptions, we state them. Where data is incomplete, we say so.

Data Sources

Our analysis draws from multiple data streams, refreshed on rolling schedules:

Data refresh cadence varies by source — some monthly, some quarterly, some annually. The "Data updated" timestamp on each page reflects the most recent refresh for that geography.

Key Metrics Explained

Rent Burden

The percentage of household income spent on rent. This is the core affordability metric.

We calculate this at the ZIP level using median rent and median household income data. A rent burden above 30% is generally considered "cost-burdened" by federal standards.

Rent Stress (Metro/State Level)

A rent-to-income ratio calculated at metro or state aggregation. Similar to rent burden but applied to larger geographies for market-level comparisons.

Crash Risk Index

A composite score (0–10) measuring downward price pressure in a metro area. Components include:

• Year-over-year price change
• Year-over-year rent change
• Percentage of listings selling below asking price
• Days on market

Higher scores indicate greater risk of price correction. A score of 10/10 means all indicators point to significant downward pressure.

Landlord-Friendliness Index

A state-level rating (1–5) based on regulatory factors that affect rental property operations:

This index describes regulatory environment, not market quality. A 1/5 state can still be profitable; a 5/5 state can still have high vacancies. Context matters.

Repit School Score

A score from 0 to 100 for every public K-12 school, built mainly on results. Families choosing a school — public, charter or magnet — should be able to see what each one delivers.

Achievement and growth (grades 3–8) come from the Stanford Education Data Archive, which converts every state’s test results to a common scale linked to the National Assessment of Educational Progress. Achievement is how many grade levels above or below the national average students score; growth (the “learning rate”) is how many grade levels students gain per school year, where 1.0 is national-average growth. SEDA’s estimates are statistically shrunk toward the mean for small schools, which limits noise. Growth matters because it reflects what a school adds, not just who it enrolls.

The weighted blend is ranked against every rated school in the same state and scaled so that roughly the top 12% earn an A+, the next 11% an A, and so on down to D.

High schools (SEDA covers grades 3–8 only) are rated on the share of students proficient in reading and math on the state test (U.S. Department of Education EDFacts; the most recent nationally published school-level year, 2018–19 or 2020–21, whichever has more schools reporting in the state) and on the four-year adjusted cohort graduation rate, which counts only regular diplomas — not GEDs or certificates of attendance. Newer high school results are published state by state; we are adding them.

Hidden gems are schools where at least half the students are low-income, yet students grow faster than the national average — with statistical confidence in 2022–25 (growth minus 1.645 standard errors still at or above 1.0), above average in 2009–19 as well, and based on at least 300 tests.

Who the school serves (context, not scored): the share of low-income students — students directly certified for free meals because their families receive SNAP or similar aid, or, in states that don’t report direct certification, students eligible for free or reduced-price lunch (NCES Common Core of Data) — and how the school’s results compare with schools in the same state serving a similar share of low-income students.

When results are missing (very small schools, pre-K and K-2 schools, some continuation and alternative schools), the school gets a small, capped rating (no higher than C+) based on resources only, and the page states that it is not a fair comparison with schools rated on results.

Limits: achievement reflects the students a school serves as well as the school. Growth is a better guide to school quality but is still estimated from public, aggregate data, not student-level records.

Repit College Ratings

Every college gets two ratings:

• Overall grade (traditional) — compares colleges the conventional way, across the five parts below.
• Repit Outcome Rating — asks what students actually earn: 60% wage-adjusted median earnings 10 years after enrolling, 25% graduation rate and 15% first-year retention. Colleges that report no earnings are not given an Outcome Rating.

The overall grade’s five parts:

How earnings are normalized: we take the College Scorecard median earnings of students 10 years after they enrolled, divide by the square root of their state’s wage index (the state’s median household income relative to the national median, from the Census Bureau’s American Community Survey) — a half-strength adjustment, because many graduates work in their college’s state and many move — and rank the result against colleges of the same type, using the Department of Education’s “predominant degree” grouping — certificate, associate, bachelor’s or graduate (for example medical and health-science universities) — with adult-serving colleges (most students 25 or older, who often had careers before enrolling) ranked separately within each group. That way a college in a low-wage state is not penalized for local pay, a cosmetology program is not compared with a research university, and a medical school is not compared with an undergraduate college.

Debt (median federal loan debt at graduation) is shown on every college page but is not part of either rating.

Earnings by major: for every college and major with reported results, we show median pay one and four years after graduation, the national median for that major, and the program’s rank among all colleges reporting that major at the same degree level — both by raw pay and by wage-adjusted pay (pay divided by the square root of the state’s wage index). Each program has its own page, for example Computer Science at The University of Texas at Austin.

Limits: Scorecard earnings cover students who received federal financial aid and were working, and include students who did not finish. Earnings reflect who enrolls as well as what a college adds.

Investability Rating

A star rating (1–5) reflecting overall investment potential at the ZIP level. Weighted inputs include:

• Cap rate estimates
• Vacancy rates
• Appreciation trends
• Rent-to-price ratio
• Crime (inverse)
• Landlord-friendliness

Weights are calibrated based on common investor priorities. This is a starting point for research, not a buy signal.

Geographic Coverage

• ZIP codes: 25,000+ with sufficient data
• Cities: All incorporated cities with available data
• Counties: 3,000+
• Metros: 894 metropolitan statistical areas
• States: All 50 states + District of Columbia

Coverage depends on data availability. Some rural ZIPs and smaller geographies may have incomplete metrics.

Limitations

How to Cite This Data

Journalists, researchers, and analysts may cite this analysis as:

For press inquiries or data requests, contact us at [email protected].