Where the data comes from

Every line in bondpassed comes from a public record and links to it. Below, for each source: how often it changes, what it shows and what it does not prove.

Check it against your own region

Open a sample (Dallas) Start free trial

Texas only. Everything within 75 miles of your office. No account needed.

This week's records: TABS registrations listed through Nov 6, 2026. All dates

Records as of this week's run

From the engine's run for the week ending Nov 5, 2026. Lists go out every Monday.

TDLR TABS: Registrations listed through Nov 6, 2026
Texas Bond Review Board: Checked Nov 7, 2026. Last changed by the Board on Nov 6, 2026
TBAE firm roster: Downloaded Nov 6, 2026
NCES district locations and Census Gazetteer: Annual files: NCES 2024-25 and the Census Gazetteer 2025

The sources

TDLR TABS (Texas Architectural Barriers System)

Texas Department of Licensing and Regulation (TDLR). The state's registry of construction projects.

Where: TABS project search. Each project we list links to its own page there, like TABS2027004959. TDLR's fact sheet: EAB At a Glance (PDF).
How often: TABS's public search is live; we read it weekly. A project page we read more than 6 days ago is read again, so a design firm added after registration shows up. This week: listed through Nov 6, 2026.
What it shows: Project name and number, owner as filed, location, type of work, estimated cost as entered, type of funds, status, and the design firm as filed.
What it does not prove: It is not a contract or a selection date. Texas law has the design professional file plans with TDLR within 20 working days of issuing them, and before a permit or construction (Gov't Code §469.102), so by the time a project is on TABS it usually names a design firm.

"Design firm" is not always an architect. The field can name an engineer, a landscape architect or a design-builder. We add "registered with TBAE to offer architecture" only when TBAE's own search ticks Architect for that firm.

The cost is what the filer typed. TABS2027005096 lists $2,000,000,000 for an elementary school. We never print an estimated cost of $500,000,000 or more, or one far too high for the square feet entered. Every cost we show reads "estimated cost entered on TABS".

"Not Assigned" does not mean no architect. A new registration may not list its design firm yet. We mark those "No design firm listed yet" and read the page again later.

Small jobs can be missing. Projects under $50,000 don't have to register, though they may (TDLR). We count building projects of $250,000 or more.

Owner names are free text. TABS2027004959 is filed by "COPPEL I S D". We show the owner we matched (Coppell ISD) next to the name as filed.

Texas Bond Review Board: Local Debt Bond Election Results

Published on data.texas.gov. The dataset's license is Public Domain.

Where: The dataset on data.texas.gov. Each bond item links to its own row, like Karnes City ISD, Prop A, Nov 3, 2026.
How often: Updates irregularly. We checked it Nov 7, 2026; the Board last changed it Nov 6, 2026.
What it shows: Issuer, election date, proposition, amount, purpose as written, and result.
What it does not prove: A blank result means not posted yet, not failed. As of our check on Nov 7, 2026, bonds from the Nov 3, 2026 election still had no result from the Board. It doesn't say when results will arrive.

On the ballot is not approved. See how we read bond results.

Its vote counts are not reliable. Karnes City ISD's 2023 row repeats its 2014 votes exactly (both rows), so we never print votes from this file.

The purpose is cut at 255 characters. We show it as written and mark where it was cut. The file lists no projects beyond that text.

County election offices (Harris, Tarrant, Bexar and Comal)

Each county's own results report for the election.

Where: Harris County Clerk
Tarrant County Elections
Bexar County Elections
Comal County, via Enhanced Voting
How often: Election night, then the official canvass. See when a result becomes official.
What it shows: Votes for and against each proposition.
What it does not prove: Election-night counts are unofficial. We use them only to take out a bond that lost in the county's unofficial count, never to say a bond passed.

We read 4 counties so far. Dallas County's results site turns away our scripts, and the rest aren't wired yet. Elsewhere a bond stays at "no official result yet" until the Board or the county posts one.

TBAE firm roster

Texas Board of Architectural Examiners (TBAE).

Where: Business rosters (roster download, Excel) and TBAE Business Search, which shows the Architect tick.
How often: TBAE says its lists are updated daily. We downloaded it Nov 6, 2026.
What it shows: Registered firms: Reg ID (BR- number), name, city, state, ZIP and status. Business Search shows whether a firm is registered for architecture, interior design or landscape architecture.
What it does not prove: Missing is not "never registered". A firm drops off the roster when its registration expires.

No people, no emails. The roster lists firms only.

TABS and TBAE share no ID. We tie a TABS design firm to a TBAE registration only through a reviewed match or the exact name and city. A near match is never printed: the firm then shows as filed, with no BR number.

NCES school district locations 2024-25

National Center for Education Statistics (NCES), EDGE program. NCES files are in the public domain.

Where: EDGE school and district locations (The 2024-25 district file we use (ZIP)).
How often: Annual. We use the 2024-25 file.
What it shows: Each public school district's office address and map point.
What it does not prove: The office, not the district's boundary. We measure to the district office. A district whose office is more than 75 miles from you is left out, even if some of its schools are closer.

Some points are wrong. Corsicana ISD's point sits in Dallas County. When a district's point is more than 15 miles from both its own ZIP and its own city, we use the city's Census point instead.

Census Gazetteer 2025

U.S. Census Bureau. Reference points for ZIP code areas, cities and counties.

Where: Gazetteer files.
How often: Annual. We use the 2025 files.
What it shows: A representative point for every ZIP code area (ZCTA), city and county. We measure distances from these.
What it does not prove: Straight-line miles, not drive time.

A point, not your door. Your region is measured from your ZIP's or city's Census point, not your street address, so a project near the 75-mile edge could fall on either side. A city's point is the Census point, not city hall; a county's is its center, not the county seat.

How we read bond results

We say a bond carried only when the Texas Bond Review Board marks it Carried ("carried per the Texas Bond Review Board") or the county posts its official canvass ("per Harris County's official canvass", for example).
Until then it reads "no official result yet". Your list, email and export leave those bonds out and show one line instead: how many bond elections await an official result, with a link to this page.
If a bond lost in a county's unofficial count, we take it out.
For the Nov 3, 2026 election, counties and districts canvass between Nov 6 and Nov 17, 2026; results are official after the canvass (Election Code §67.003, §65.051(a-1)).

What's covered

Texas only. No other state yet. TDLR TABS and the Texas Bond Review Board cover the whole state.
Your region: everything within 75 miles (straight-line) of your office.
Building projects: TABS registrations marked public funds, by an owner we can match to a public body, with an estimated cost entered on TABS of $250,000 or more, not civil or site work, and not finished. A private owner marked public funds is left out. A program filed once per campus counts as one project.
Bonds: school districts, cities and counties. Community-college and hospital-district bonds aren't included yet; water-district bonds are left out on purpose.
County vote counts: Harris, Tarrant, Bexar and Comal only.

How we check

30/30 random findings reproduced from live records
400/400 facts matched the record
0 false findings
120/120 source links worked

Before launch, on Nov 6, 2026, a separate check, with code that shares nothing with our engine, took 30 findings at random from the Austin, Dallas and Tyler regions and re-derived each fact from the official records, fetched live; all 30 matched. To test the check itself, errors were planted in copies of findings; it caught every one.

Every week:

Each fact is compared with the stored copy of its record before it goes out. A mismatch drops the item.
Records are tied together only through reviewed tables or an exact name plus a place check, never a near match. Anything unresolved stays out.
A run is published only after the engine's tests pass, including a case for every confirmed wrong record.

Wrong record?

Every item in your list has a "Wrong record?" link. A person checks your report against the state record, usually the same day.

If our line is wrong, we fix the engine and add the case to the tests every run must pass. Leave your email and we'll tell you when it's fixed.

We show what each record says, as filed, typos included. If you think the record itself is wrong, report it and we'll check it against the source.

Using these records

TDLR and TBAE let their information be copied when the copy names the agency, its web address and the date, with a statement that it isn't endorsed by the State (TDLR's terms, TBAE's terms). Every line we show links to its record, and the dates are above.

Not affiliated with or endorsed by the State of Texas, TDLR, the Texas Bond Review Board or TBAE.

Check your region Open a sample (Dallas)