Every figure here ships with the script that reproduces it.

Who publishes this

The field studies are written and counted by Chris Klebl and published by Sterling Digital Partners, an independent software studio and the company behind Glove. He is the author of record for all five studies, the person who ran each count, and the contact for corrections. Reach him at [email protected].

There is no newsroom here and no research staff. That is worth stating plainly, because it is the reason the method is published in as much detail as it is: the work has to be checkable by you rather than vouched for by an institution.

What “counted” means on this site

It is a narrow word here and it carries four commitments.

A published script produces every figure. Each study ships a count_*.py that uses the Python standard library only — no dependencies, no API key, no account. It pulls the same public source the study names, prints every figure the study quotes, and writes them to a CSV. If a number appears on a study page and not in that CSV, the page says so and says where it came from instead.

The source is named, dated and linked. Not “county records” but the actual file or service, the date the file carried, and the date we read it. Where a county publishes no dated snapshot — which is common, and is itself one of the findings — the study says the figure cannot be reproduced for a past date by anyone, including us.

Extracts are reconciled before anything is counted. Every pull is checked against the source’s own count of itself and de-duplicated on its identifier. One county’s service returns 57 rows that do not exist; the only way to know that is to check, every time.

Findings that fail are published as findings. A pattern found in one market and tested in the next either reproduces or it does not, and both outcomes get printed. “North Carolina has no assessment cap” is a finding of study No. 03 precisely because the Florida measurement had nothing to bite on there.

What this research is not

It is not a national estimate. Eight counties in four states is eight counties in four states, and every study says so in its own words. It is not a sample of any state either — the markets are chosen because Glove covers them, which is a commercial reason and not a statistical one.

It is not an audit of any assessor. Most of what these studies count is a record working exactly as its statute instructs. Where a figure looks like an error it is usually a design, and the studies say which.

And it is not disinterested. Sterling Digital Partners sells Glove, a tool that pulls property records. The research exists partly because it is the best demonstration of why that job is harder than it looks. Every figure is still counted the same way and published with the means to check it, which is the only answer to that conflict that actually helps you.

Corrections

If a count looks wrong, send the metric name and the public source to [email protected]. We re-run it, correct it, and re-date the page. Corrections are made in place with the change stated rather than quietly; the records directory keeps a public change log that records open questions and rejected sources as well as resolved ones.

Two corrections already made this way are described in the studies themselves — a Miami example that turned out to be four condominium wing labels rather than four directional collisions, and a Manatee half-study rebuilt onto static source files after the original source proved to be regenerated nightly. Both are on the record because a method that only reports its successes is not a method.

Using this work

The counts CSVs, the aggregate dataset and both scripts in each study are published under CC BY 4.0. Quote the figures, republish the data, build on the scripts — commercially is fine — with attribution to Chris Klebl and Sterling Digital Partners and a link to the study you took it from.

Every chart carries an embed snippet with both a PNG and an SVG version, sized and attributed. Journalists and analysts are welcome to the underlying files without asking; if you want a cut of the data that is not published, or a market counted that is not yet in the series, email and ask.

The series so far

No. 01 — Sarasota and Manatee counties, Florida. No. 02 — Miami-Dade County, Florida. No. 03 — Mecklenburg County, North Carolina. No. 04 — Harris County, Texas. No. 05 — Charleston, Berkeley and Dorchester counties, South Carolina. All 330 figures from all five are collected, each with its source and date, on the statistics page.

The index carries an RSS feed and a per-market notification list for the studies still to come.