15 public sources · 21 published measures · nothing behind a login
“Is this accurate?” is not a question you should have to take on trust. Below is every government file this site is built from, with a download link, and every published figure tied to the exact column it came from, the arithmetic done to it, and the test that recomputes it from the state’s own file. Nothing here requires a login, a licence or a records request — you can repeat this work from scratch.
Texas Education Agency · fiscal 2009–2025, every Texas school district
revenue, spending by object and by function, debt service, capital outlay, payroll, enrolment.
Districts file this with TEA and it is corrected over time. The newest year has had the least time to be corrected.
Texas Education Agency · 2009–2024
student demographics, teacher counts, turnover, experience and salary, class size, attendance, graduation, STAAR at three bars, county and region.
Texas Education Agency · school years 2024 and 2025
results by student group, including economically disadvantaged students specifically.
District-level aggregates are also downloadable from the Texas Assessment Research Portal at https://txresearchportal.com/ .
Texas Education Agency with the Texas Comptroller · certified property values tax years 2015–2025; adopted tax rates 2005-06 to 2023-24
taxable value, M&O and I&S tax rates.
Tax year N is TEA fiscal year N+1. Figures are withheld where two independent estimates of the tax base disagree by >25%.
Texas Education Agency · fiscal 1994–2026
how much local tax revenue each district sends back to the state.
Texas Bond Review Board · 4,992 decided propositions, 1958–2026
what school debt was asked FOR, the amount, and whether voters carried or defeated it.
Corrected 2026-08-11. This entry previously credited the file to compiled county returns from the Secretary of State and stated that no single agency publishes school bond elections statewide. Both were wrong: the Bond Review Board publishes all of them, back to 1958, on the state's own open-data portal, and always had. The site had been shipping a municipal-advisory vendor's Excel export of that same file — two years stale, 404 propositions short, and carrying spreadsheet subtotal rows. verify_sources.py did not catch it because the Secretary of State URL returns 200; it proved the link was alive, not that it was the right link. It now checks attribution as well. Every row carries the Board's own provenance mark in a Source column (TBR, Issuer, OAG, TSB, BB, Other). This is still the only layer joined on district NAME rather than TEA number; scripts/audit_bond_match.py prints the whole join and fails the build if a shared name is resolved without the county agreeing. Two companion files from the former vendor carry a private CRM and are deliberately never ingested — going first-party removes the temptation permanently.
Texas Education Agency · 2025 ratings for 9,084 campuses and 1,208 districts, with enrolment and student need; the workbook also carries a longitudinal sheet back to 2011
the A–F rating the state gives each campus and each district, and the component scores behind it.
The only campus-level source on this site. Two distinctions survive into every downstream count: 'Not Rated' is missing data rather than failure (525 campuses, all excluded), and Alternative Education Accountability campuses are rated on a different scale and are excluded from the headline with the including-them figure published beside it. No campus is ranked against a campus in another district — the published claim is about the gap INSIDE a district, which is what a district rating conceals.
Texas Bond Review Board · fiscal 2005–2025 actual, plus the amortisation schedule for debt already sold out to 2061; 967 districts
principal and interest still owed, split into current interest bonds and capital appreciation bonds, and the year the borrowing clears.
The Board's own issuer index and per-issuer series, taken first-party. A commercial aggregator republishes the same data and is how it was found; it is deliberately not the source, because a re-publisher breaks the provenance chain at its first link. Joined to TEA numbers by name plus county via scripts/district_match.py — the index lists three issuers under two names each, so it is deduplicated by the Board's id rather than by name, which is what stops one district's debt being counted twice. Excludes obligations under one year, commercial paper, and special obligations not needing Attorney General approval, per the Board's scope note.
US Census Bureau (public domain) · 2025 boundaries, 1,016 districts
district geographic boundaries.
Charter districts have no attendance boundary, so ~8% of students are not on the map by construction.
US Census Bureau (public domain) · fiscal 2024, every U.S. public school system
national per-district revenue, spending, and per-pupil current spending on one federal ruler.
Current spending is the Census's own concept: it excludes construction, land, debt, and spending for students outside a district's fall membership, so it is deliberately never mixed with the TEA all-funds figures on this site. Dollars in the file are thousands; per-pupil columns are whole dollars. Fiscal 2024, one year behind the TEA data here.
National Center for Education Statistics · fiscal 2024, the 50 states, DC and 5 territories
state-level per-pupil current expenditure and the Texas state ranking.
PPE15 divides current expenditure (NCE13) by average daily attendance (ADA) — an identity the build asserts for every state. Territories are excluded from the ranking; the artifact says so.
National Center for Education Statistics · SY 2024-25, every U.S. local education agency
the Census-to-TEA join: ST_LEAID ('TX-057905') to NCES LEAID (4816230).
The NCES district id is NOT derivable from the TEA number by arithmetic (Dallas is 4816230, not 48+057905); this directory is the only published bridge. Keyed by number, so the eleven Texas district names shared by two districts each cannot cross-match. Every Texas F-33 row resolves through it, and the build refuses to write the artifact if the accounting does not balance.
Universal Service Administrative Company · funding years 2016–2026, every Texas funding request
federal E-Rate commitments and authorized disbursements for school connectivity.
Sums on this site cover FRNs with status 'Funded' ONLY: a Pending or Cancelled row's funding_commitment_request is the amount the applicant ASKED for, not a commitment, and summing every status inflates a funding year by hundreds of millions of never-granted requests. E-Rate discounts are mostly paid by USAC straight to vendors, so this money appears nowhere in TEA's books.
Universal Service Administrative Company · every Texas E-Rate entity (applicants and their schools)
the applicant-to-district join: billed entity number to TEA district number.
Self-reported in EPC and sparse — Dallas ISD's own row carries no state code at all. Resolution runs three roads (state code verified against the crosswalk, unanimous campus-code prefix of child schools, audited name+county) and REFUSES where they disagree; the one conflict on record is the two Dawson ISDs, a twin name. Charter networks spanning several TEA districts under one applicant number are never attributed to one of them.
US Bureau of Labor Statistics · 2009–2025
converting past dollars to constant 2024 dollars.
If a figure appears on this site and is not in this table, that is a bug — please report it.
| Measure | Source | Columns used | How it is computed | Re-derived by |
|---|---|---|---|---|
| Spending per student /district/{n}/summary |
Summarized PEIMS Actual Financial Data | all_funds_total_operating_expenditures_by_obj all_funds_total_debt_service_expend_by_obj fall_survey_enrollment |
Operating spend plus debt service, divided by fall enrolment. Composed, never subtracted — TEA's operating total excludes debt service. | test_statewide_debt_total_recomputes_from_source |
| Share of the operating dollar reaching a classroom /trends/texas |
Summarized PEIMS Actual Financial Data | all_funds_instruction_transfer_expend_fct11_95 all_funds_total_operating_expenditures_by_obj |
Function 11–95 (instruction) as a percentage of total operating expenditure, summed across districts. | test_instruction_share_recomputes_from_source |
| Debt service per student /district/{n}/forensics |
Summarized PEIMS Actual Financial Data | all_funds_total_debt_service_expend_by_obj fall_survey_enrollment |
Debt service divided by enrolment, deflated to constant 2024 dollars with CPI-U. | test_debt_per_student_recomputes_from_source |
| Security and monitoring per student /trends/texas |
Summarized PEIMS Actual Financial Data | all_funds_security_monitoring_service_expend_fct52 fall_survey_enrollment |
Function 52 divided by enrolment, in constant 2024 dollars. | test_security_per_student_recomputes_from_source |
| Operating revenue minus operating spending /trends/texas |
Summarized PEIMS Actual Financial Data | all_funds_total_operating_revenue all_funds_total_operating_expenditures_by_obj |
Operating revenue against operating spending ONLY. Debt service is excluded from the cost side and the I&S debt levy (all_funds_other_revenue) from the income side — it tracks debt service to within 15% every year, so counting it as operating income understates deficits badly. | test_the_operating_deficit_uses_operating_revenue_only |
| Who pays: local, state and federal shares /district/{n}/economics |
Summarized PEIMS Actual Financial Data | all_funds_local_tax_revenue_from_m_o all_funds_state_revenue all_funds_federal_revenue |
Shares of GROSS local collections. TEA reports M&O revenue net of recapture, which makes a paying district look less locally funded than it is. | test_local_revenue_is_gross_not_net_of_recapture |
| School tax on a $300,000 home /district/{n}/economics |
School District Property Values and Tax Rates | adopted M&O rate adopted I&S rate |
Adopted total rate applied to a $300,000 taxable value. Null for charters (no property tax) and where two independent estimates of the tax base disagree by more than 25%. | test_the_revenue_shares_are_a_whole_dollar |
| Recapture paid to the state /district/{n}/economics |
Excess Local Revenue (recapture) paid by district | recapture paid by district | As published by TEA, divided by enrolment. | test_local_revenue_is_gross_not_net_of_recapture |
| Campuses rated D or F inside a district rated A or B /campuses/texas |
Enhanced Statewide Summary — A–F accountability ratings | Overall Rating (campus) Overall Rating (district) Number of Students Alternative Education Accountability |
Campus rating compared with the rating of its own district. Unrated campuses excluded; alternative-education campuses excluded from the headline and reported separately. | test_the_headline_is_the_sum_of_its_own_campuses |
| Debt still owed — principal plus interest not yet paid /district/{n}/debt |
Local government debt outstanding — school districts | CIBPrincipalOutstanding CIBInterestOutstanding CABPrincipalOutstanding CABInterestOutstanding |
Summed per district for fiscal 2025, the Board's most recent completed year. Years after 2025 are the amortisation schedule for debt already sold and are never added to it. | test_the_statewide_total_is_the_sum_of_its_districts |
| Interest deferred to maturity on capital appreciation bonds /debt/texas |
Local government debt outstanding — school districts | CABPrincipalOutstanding CABInterestOutstanding |
Reported as an absolute stock for the current year. Dollars repaid per dollar borrowed is reported at each district's PEAK year only — on a shrinking balance the ratio rises on its own and measures nothing. | test_the_repayment_ratio_is_never_taken_on_a_residual_balance |
| Every bond proposition and how the vote went /district/{n}/bonds |
Texas school district bond election results | Issuer County Elect. Date $ Amount Purpose Description Result Votes For Votes Against |
Matched to TEA district numbers by stemmed name plus county — Texas has thirteen pairs of districts sharing a name. 100% matched, 0 resolved without the county agreeing. | test_published_bond_data_matched_everything_and_says_how |
| Results against what student need predicts /district/{n}/outcomes |
Snapshot — District and Charter Detail | test_all_meets pct_econ_disadv pct_emergent_bilingual pct_special_ed |
Least squares of the Meets rate on three need measures; the published figure is the residual. The model explains about half the spread between districts — it is an estimate with real uncertainty, not a fact. | test_percentiles_are_in_range_and_mean_the_same_thing_everywhere |
| How a district's low-income students do /district/{n}/equity |
STAAR Aggregate District-Level Results | economically disadvantaged group, Meets grade level | Reported at the Meets bar and benchmarked against poor students statewide — never as a gap, which correlates about zero with how poor students actually do. | tests/test_api.py |
| District boundaries on the map /district-geo |
TIGER/Line Unified School Districts, Texas | GEOID geometry |
Joined to TEA district numbers. Geolocation is resolved entirely in the browser; a visitor's location is never sent anywhere. | tests/test_static_pages.py |
| A district against every U.S. district /district/{n}/national |
Annual Survey of School System Finances (F-33), district file, fiscal 2024 | PPCSTOT ENROLL NCESID |
The Census's own per-pupil current spending figure, placed among the U.S. districts with 500+ students (the pool size travels in the artifact, never in prose). The percentile is the share of ranked districts spending strictly less. Never mixed with TEA figures: different fiscal year (2024) and a narrower ruler (no construction, no debt). | test_the_district_percentile_rederives_from_the_census_file |
| Texas among the 50 states and DC /national/texas |
National Public Education Financial Survey (NPEFS), state file, fiscal 2024 | PPE15 NCE13 ADA FIPS |
States ranked by NCES's published per-pupil current expenditure, which is current expenditure divided by average daily attendance — an identity asserted for every state at build time. Territories excluded, and said so. | test_the_state_rank_rederives_from_the_committed_npefs_file |
| The Census-to-TEA district join /district/{n}/national |
CCD Local Education Agency Directory, SY 2024-25 (v.1a) | LEAID ST_LEAID FIPST |
NCES ids are not derivable from TEA numbers by arithmetic, so every join goes through the CCD directory's ST_LEAID column, keyed by number — shared district names cannot cross-match. Every Texas F-33 row resolves, or the build refuses to write the artifact. | test_the_bridge_is_keyed_by_number_not_name |
| Federal E-Rate money for a district's connectivity /district/{n}/erate |
E-Rate Request for Discount on Services: FRN Status (FCC Form 471) | funding_commitment_request total_authorized_disbursement form_471_frn_status_name ben funding_year |
Funded-status FRNs only, summed per funding year for the district's own billed entity number(s). Consortium money is counted statewide and never attributed to member districts. Years still inside their invoicing window are flagged open; a drawn percentage there is a floor. | test_the_statewide_series_rederives_funded_only_from_the_raw_file |
| The E-Rate applicant-to-district join /district/{n}/erate |
E-Rate Supplemental Entity Information | entity_number state_lea_code state_school_code parent_entity_number physical_county |
Three roads — verified state code, unanimous campus-code prefix of child schools, audited name+county — that must agree; a disagreement is refused, never guessed. The accounting (matched + refused + unmatched = all applicants) is asserted at build time. | test_the_match_accounting_balances_and_refusals_are_recorded |
| Constant 2024 dollars /economics/texas |
CPI-U, annual average | CPI-U annual average | One price base for the whole site. The trend layer reuses the economics layer's factors rather than defining a second. | test_debt_per_student_recomputes_from_source |
git clone https://github.com/GusAI40/texas-isd-finances cd texas-isd-finances python -m pytest tests/test_provenance.py # re-derive every headline from source python scripts/verify_artifacts.py # prove the served files match a fresh build
The financial file every headline is built from, as captured. If the state restates a year this hash stops matching and the tests fail rather than the change being absorbed silently.
Machine-readable: everything on this page is served as JSON at /provenance, generated from the same register, so the page and the API cannot disagree. Data disclaimer: an independent transparency tool presenting Texas Education Agency, Comptroller and US Census data for informational purposes only. Not an official TEA product, not an audit, and not financial, legal or investment advice.