Data

Downloads

Free to reuse with attribution. Carry the reporting period when you republish a figure. The JSON files the pages read are open too, with no key, and their paths are stable. data/metrics.json lists every measure and its unit, and data/sources.json lists every source, period and limit.

Sources

Process

Each update runs six steps.

1. Collect

   Every morning the site downloads eight public sources from New York City Public Schools, NYC Open Data and the City's geocoder. The reporting period is read from inside each file.

2. Match

   Every row is matched to a school by its DBN, the ID every City school has. Never by name, because each file spells names differently.

3. Check

   Each figure keeps its year, unit and status. A figure the source did not publish stays blank. The build stops if a source changes shape or loses schools, rows or locations.

4. Band

   For about one figure in six, the City publishes a score from 1 to 5 against a group of similar schools. A profile colors four bands from that score, 4 and up, 3 to 4, 2 to 3 and under 2, with the score inside. No score, no color. That grouping is the site's only judgment.

5. Count

   For each figure, the build counts the schools of the same type by their published value, in 24 equal steps, for the two newest years. A profile draws where its school falls among them. The counts name no school and make no score.

6. Place

   Each school gets a point from the City's school point file, then the high school directory, then the geocoder. A school none of them has stays off the map.

7. Publish

   The same checked tables write every page and every download, so they cannot disagree. A late source puts a warning on every page.

Figures

• A band describes one measure in one year. It is not a rating of the school, and the site has no overall score or ranking. Each school is scored against its own group, so Compare shows figures without bands.
• The bars beside a figure count schools of the same type, as the City types them in its reports, by their value that year. Schools the City publishes only as "Above 95%" have a bar of their own.
• Impact and Performance are the City's scores from 0 to 1. Performance compares a school's results with the citywide average for its type. Impact compares them with how similar students do across the city. The City centers both on 0.50 within each school type and does not cap them at 1 in its download.
• The City rates three areas: Instruction and Performance, Safety and School Climate, and Relationships with Families. The first digit of the score, from 1.00 to 4.99, sets the word. The site shows both and adds no band.
• Survey results are the share of positive answers, beside the City average for the same report. Response rates show how many surveys came back.
• Advanced course shares by race are each group's share of the students in those courses.
• A profile shows each source's newest year for that school, named once beside the source. A measure with no figure that year shows as not reported, and its earlier years are in the opened row. A measure the City no longer publishes shows its last year beside the figure.
• A blank is never a zero. Not reported means schools of the same type publish the measure and this one has no figure. Withheld means the group is too small to publish. A measure that cannot apply, such as an 8th-grade figure at a K-5 school, is left off the profile and reads Does not apply in Compare. Compare leaves off a row that applies to none of the chosen schools.
• Compare holds five schools. A measure the City publishes separately for each report type, such as attendance for elementary and high schools, is one row there, and each school reads its own report's figure.
• Where no charter school has a figure for a measure, a charter school's profile says so.
• A school is open if it is in the latest enrollment snapshot or a current directory. No source publishes a status, so a new school can be labeled wrongly.
• Admissions figures are per program and describe the Fall 2025 season only.
• Schools report in different years, and each comparison row shows its year.
• 2019-20 and 2020-21 reflect the pandemic, when testing was canceled or thin.
• Some measures changed definition in 2018-19 and appear as separate series.
• A school with grades 6 to 12 files two reports, shown as two rows.
• A percentage of 12 students is less stable than one of 900. Group sizes are shown where published.
• Sorting the comparison table orders one measure. It is not a ranking.

Limits

• The site has no overall scores, ranks or recommendations.
• It has no address search, school zones or commute times.
• SHSAT figures, specialized high school admissions, suspensions and school budgets are not included.
• Ratings, Impact and Performance cover 2023-24 and 2024-25. District 75 and early childhood schools do not receive them.
• The City counts a student as receiving all recommended special education programs when the student's class schedule matches the IEP.
• Charter schools outside the City's own reporting are not included.