About OpenSchoolData

Why federal school data deserves a public-facing home

The U.S. Department of Education’s National Center for Education Statistics (NCES) maintains the Common Core of Data (CCD), the federal dataset covering every public school, every school district, and every state education system in the United States. The CCD includes enrollment, staffing, demographics, fiscal data, and a separate set of files for student outcomes (test scores, graduation rates, attendance). The Urban Institute distributes a public-friendly API on top of this federal data, and OpenSchoolData reads through that API.

The presentation problem is the usual one: the federal data is rigorous and comprehensive, but the official portal is built for state education-agency researchers. A parent evaluating a school, a board candidate researching local district outcomes, or a journalist comparing graduation rates needs a presentation layer that makes the same data answerable per-school, per-district, and per-state without a multi-step data download.

OpenSchoolData builds that layer. Every school page consolidates the NCES record for that school: enrollment, demographic breakdown, free-and-reduced-lunch share, teacher-student ratio, graduation rate (for high schools), and the multi-year trend. Every district page rolls up the schools in the district. Every state page rolls up to the state-level distribution. The data is the NCES data that has always been public; the value the site adds is the navigation.

How the pipeline pulls NCES data

The pipeline pulls from the Urban Institute Education Data Portal on a four-month cadence aligned to each NCES release cycle. The CCD vintage covers the previous-prior school year, so the September release typically carries data from the school year that ended 16-18 months earlier. The site stamps the as-of date on every value so readers know which CCD vintage they are reading.

A practical detail: NCES outcome data (test scores, graduation rates) is collected through the EDFacts data submission, which is a separate file from the core CCD enrollment data and is on a slightly different release schedule. The site combines both where coverage is available; in states where EDFacts data is sparse or suppressed for privacy at small-school levels, the pages flag the gap rather than imputing values from peer schools.

Where K-12 outcomes data has caveats

Three caveats. First, NCES suppresses small-cell data to protect student privacy — a graduation rate for a class of 12 students would identify individuals, so the figure is suppressed. The site shows suppressed values explicitly rather than imputing. For very small rural schools, several outcome fields can be suppressed and the page leans on the comparable district-level figures instead.

Second, NCES data lags the school year by 1-2 years. A school that opened in fall 2024 may not appear in NCES data until the 2026 release. Schools that closed or reorganized appear in historical data even after they no longer operate; the pages flag known closures where the state agency has updated the record.

Third, charter schools and non-traditional school types (special-education centers, alternative schools, hybrid online schools) are in the dataset but their outcome metrics are not always directly comparable to traditional schools. The pages note the school type explicitly and provide guidance on which comparisons are meaningful for each type. Every numeric value links back to the originating NCES record for verification.