Where the knowledge base's facts come from
Every history and reference page on this site was written from a fact sheet, and every sheet was checked twice against the documents it cites. This page shows the method, lists the 254 archives and agencies the 1,139 documents come from, and publishes the claims themselves as data: 2,229 claims across 45 pages, each with its status, its quotation and the documents behind it, in one JSON file per page. A copy of every cited document was also pulled down and kept with the site's records, so a page can be checked against the very document it came from long after a link has died.
Last updated September 2026 · Counts are read from the fact sheets and the claims data on the date given · No prices on this page; the Estimate Builder carries those · Part of the knowledge base
How a fact gets onto a page
Each history and reference page in this library began as a fact sheet: a list of claims, each with the document it came from and a short quotation from that document, written by a research pass that read the sources directly. A second pass re-opened every source and tried to refute the sheet, correcting what it could not confirm and marking each claim as verified when two independent publishers state it, single when one named publisher does, or a conflict when the documents disagree. The page was then written from the surviving claims only, and a final check traced every sentence on the page back to a claim. The sheets, the verification reports and the page checks are kept with the site's source code, and the claims themselves are published here as data.
The claims as data
One file per page, in JSON, lists every claim the page rests on with its status, its quotation and the documents behind it, plus the conflicts the page prints. An index lists the files. The set covers 45 pages, 2,229 claims and 1259 source entries, and it is regenerated when a page changes.
kb-data/index.json is the index. The files below are the pages that carry a fact sheet; the site's other pages describe the company's own work and cite their few sources in place.
Where the documents live
The 1,139 distinct documents behind the sheets come from 254 web hosts and 435 named publishers. The hosts that supply six or more are listed here with what the pages take from them; the 216 that supply fewer than six between them account for 383 documents and are named on the pages that use them.
| Host | Who publishes there | Documents | Pages | What the pages take from it |
|---|---|---|---|---|
| codes.ohio.gov | Ohio Laws and Administrative Rules | 147 | 26 | the Ohio Revised Code, the Ohio Administrative Code and the building, plumbing and residential codes as Ohio adopts them |
| www2.census.gov | United States Census Bureau | 130 | 25 | the 1940 and 1950 housing volumes and the 1950 census tract bulletin for Columbus |
| archive.org | Internet Archive | 99 | 28 | the Ohio State Board of Health annual reports from 1886 on, the county histories of 1880 and 1908, Lee 1892, Hooper 1920, Studer 1873 |
| digital-collections.columbuslibrary.org | Columbus Metropolitan Library | 94 | 14 | Columbus Memory, the city engineer photographs, the Ohio postcard and photograph collections |
| columbus.gov | City of Columbus | 73 | 21 | neighborhood plans, the water and sewer divisions, Blueprint Columbus, building and zoning pages |
| dam.assets.ohio.gov | State of Ohio document store | 52 | 20 | Ohio EPA and Ohio Department of Health publications |
| epa.gov | United States Environmental Protection Agency | 35 | 9 | drinking water rules, lead and copper, PFAS, the federal drinking water inventory |
| web.archive.org | Wayback Machine | 33 | 11 | captures of city and village pages that refuse fetchers or have changed since they were read |
| up.codes | UpCodes | 28 | 7 | the model codes as adopted, where the code council refuses direct reading |
| remarkableohio.org | remarkableohio.org | 27 | 6 | |
| census.gov | United States Census Bureau | 25 | 8 | population estimates, the American Housing Survey, county and place tables |
| nps.gov | National Park Service | 23 | 8 | the National Register program and preservation briefs |
| catalog.archives.gov | National Archives Catalog | 20 | 8 | National Register of Historic Places nominations in full |
| huduser.gov | United States Department of Housing and Urban Development | 17 | 9 | housing research and the Poindexter Village case study |
| loc.gov | Library of Congress | 17 | 9 | the Sanborn fire insurance maps of Columbus, Delaware, Westerville and Worthington, and photographs |
| npgallery.nps.gov | National Park Service, NPGallery | 17 | 8 | National Register listings and nomination documents |
| auditor.franklincountyohio.gov | auditor.franklincountyohio.gov | 17 | 7 | |
| ohiohistory.org | Ohio History Connection | 16 | 5 | Ohio Memory and the historical markers |
| cpsc.gov | cpsc.gov | 14 | 7 | |
| com.ohio.gov | com.ohio.gov | 13 | 7 | |
| tcnatile.com | Tile Council of North America | 10 | 2 | the tile installation handbook |
| pubs.usgs.gov | United States Geological Survey | 10 | 8 | flood circulars and water-supply papers |
| ecfr.gov | Electronic Code of Federal Regulations | 10 | 5 | the federal rules the pages quote |
| columbus-oh.elaws.us | columbus-oh.elaws.us | 10 | 2 | |
| ohiodnr.gov | ohiodnr.gov | 8 | 3 | |
| worthington.org | worthington.org | 8 | 3 | |
| water.noaa.gov | water.noaa.gov | 8 | 1 | |
| ohiohistorycentral.org | ohiohistorycentral.org | 7 | 2 | |
| wosu.org | WOSU Public Media | 7 | 6 | Columbus Neighborhoods |
| gahanna.gov | gahanna.gov | 7 | 2 | |
| epa.ohio.gov | Ohio Environmental Protection Agency | 7 | 5 | permits, fact sheets and the PFAS action plan |
| msc.fema.gov | msc.fema.gov | 6 | 3 | |
| dublinohiousa.gov | dublinohiousa.gov | 6 | 3 | |
| apps.franklincountyauditor.com | apps.franklincountyauditor.com | 6 | 5 | |
| worthingtonmemory.org | worthingtonmemory.org | 6 | 1 | |
| codelibrary.amlegal.com | codelibrary.amlegal.com | 6 | 2 | |
| blueprintneighborhoods.com | blueprintneighborhoods.com | 6 | 2 | |
| delawarehealth.org | delawarehealth.org | 6 | 2 |
The copies kept on hand
Web pages change and archives move. So that a page can be checked or corrected years from now without depending on a link still working, a copy of every document the sheets cite was pulled down on September 6, 2026 and kept with the site's records, with a manifest recording each file's origin, size and checksum: 706 of the 1,139 documents were captured, most directly and the rest through the Wayback Machine, and 28 refused every route and are listed with the status they returned. The photographs and maps the place histories use were kept at the largest size their archives serve, with each archive's own rights statement quoted beside the file. None of those copies is published; the pages link to the archives themselves.
What the data does not contain
No claim about this company, no price and no advice to hire anyone is in any sheet; the sheets describe Ohio, Central Ohio and the documents that record them. Photographs that an archive marks as copyrighted or undetermined were recorded and not copied, and appear nowhere on the site. Where a document names a private household, the sheet keeps the place and the date and drops the name.
Sources
Every figure, date and quotation on this page comes from one of these documents, read on the date given. Where two of them disagree, the page says so.
- Internet Archive, The Internet Archive digital library, undated; read September 2026.
- Library of Congress, Sanborn Maps collection, undated; read September 2026.
- United States Census Bureau, Census of Housing and Census Tract Statistics, decennial volumes, undated; read September 2026.
- HathiTrust, HathiTrust Digital Library, undated; read September 2026.
- Columbus Metropolitan Library, Digital Collections, undated; read September 2026.
- National Archives, National Archives Catalog, National Register of Historic Places records, undated; read September 2026.
- Ohio Laws and Administrative Rules, The Ohio Revised Code and Ohio Administrative Code, undated; read September 2026.
- Internet Archive, Wayback Machine, undated; read September 2026.
Common questions
Where do the facts on this site's history and reference pages come from?
From named public documents, read directly: the Ohio State Board of Health's annual reports, the Census Bureau's housing and tract bulletins, the Sanborn fire insurance maps at the Library of Congress, National Register nominations, city plans and agency pages, county histories, and the codes as Ohio adopts them. Every page lists its own sources at the foot, and this page lists the hosts and publishers across the whole library, with a count of how many documents each supplies and how many pages use it.
What does it mean when a claim in the data is marked verified?
That two independent publishers state the same fact. A city page and a PDF from the same city count as one publisher, and a document that merely cites another does not make two. A claim marked single rests on one named publisher, and the page names that publisher in the sentence that uses it. A claim marked conflict means the documents disagree; the page prints both versions with their publishers and does not choose between them.
Can the claims behind a knowledge base page be downloaded as data?
Yes. Each page that carries a fact sheet has a JSON file listing its claims, their status, the quotation each rests on and the documents behind it, along with the conflicts the page prints. The index file lists every page and its data file with a claim count. The files are plain JSON, regenerated whenever a page changes, and free to read by any program or assistant that reads the site.
How were the documents behind the pages checked?
Twice, by separate passes. The research pass read each source and wrote the claim with a short quotation from the document. A verification pass then re-opened every source and tried to refute each claim, correcting dates and figures it could not confirm, downgrading claims whose two sources turned out to be one, and recording what it could not settle. After the page was written, a page check traced every sentence back to a claim id and re-opened the sources behind the figures a reader would most want to trust.
Why are copies of the source documents kept if the pages link to them?
Because links stop working. City pages are rewritten, archives reorganize, and some sites refuse automated readers on some days and not others. A kept copy, with its origin, size and checksum recorded in a manifest, means a page can be checked or corrected later against the very document it was written from. The copies are for that purpose only; they are not published, and the pages continue to link to the archives that hold the originals.
Why do some pages cite the Wayback Machine instead of a city or village website?
Because the site refused the reader that day or has since changed. Several city and village websites decline automated requests intermittently, and a page read through the Internet Archive's Wayback Machine is cited as that capture, with its date, so a reader can open exactly the version the fact was taken from. Where the live page and a capture differ, the page says which one it quotes.
Which archives supply the most documents to the knowledge base?
Ohio's own code publisher, the City of Columbus, the Internet Archive, the United States Environmental Protection Agency and the Wayback Machine head the list, followed by the Columbus Metropolitan Library's digital collections and the Census Bureau. The Internet Archive matters most for the nineteenth century: the State Board of Health reports, the county histories and the city histories of 1892 and 1920 are read there, page by page, and quoted from the printed text rather than from a summary.
Are the fact sheets and the data written by people or by software?
By AI assistants working from the documents, under rules that bind every sheet: a claim must quote its source, a figure must carry its table heading and its unit, a superlative appears only where a document states it, and a disagreement between documents is printed as a disagreement. The rules and the results are the product; the reader can test any sentence against the claim file and the document it names, which is the point of publishing them.
See your actual price in about two minutes
Pick your shower, choose real products with real model numbers, and watch the number move. No phone number required, and nobody will call unless you ask.