Dataset · MVS-R 34 · Reclaim City
Who owns, supplies and buys from data brokers
An open dataset of mostly US data brokers and the documented links between them. Every published link cites a regulator, court, SEC or federal procurement record, and every record has an archived copy. Built for Reclaim City, a game about taking your data back, and published for journalists, researchers and privacy advocates to check, reuse and extend.
- 970Broker records
- 106Published links
- 174Source documents
- 742Opt-out routes
As of 2026-10-24 · CC BY 4.0
Download the dataset
The complete dataset
Every broker record, every link (published or not), every source, the state registry snapshots, the research notes and the build scripts, with the README and license.
ZIP · 1.3 MBPublished export
One file: all 970 brokers with their opt-out routes, reach and hubs, and the 106 tier A links with their sources.
JSON · 471.5 KBBrokers table
One row per broker: registries, what it sells, how to opt out, whether that route was verified, and its reach.
CSV · 144.9 KBLinks table
One row per published link: who, what kind of link, to whom, the narrow claim, and the source with its archived copy.
CSV · 59.4 KBPatreon · Back the research
This research is free because patrons pay for it
Nothing on these shelves sits behind a paywall, and the datasets are free to reuse. Patreon is how it stays that way: it pays for the research, the instructors and the scholarships. A few dollars a month counts.
Methods
What counts as evidence
Every link has a tier. Only tier A is published: a regulator or court document, an SEC filing, a statutory annual report filed with a non-US securities regulator, or a US federal procurement record. A company's own statements (tier B) and reputable journalism (tier C) are kept in the full dataset as research leads, marked unpublished, not as findings. Of 158 links researched, 106 are published:
47 owns, 44 buys, 15 supplies.
- Narrow claims. Each claim says only what its source says. Allegations in a complaint are written as allegations, not findings.
- Archived and checkable. Every source has a Wayback Machine copy, so a claim can be checked even if the live page moves.
- Unknown stays unknown. A missing size is unknown, never an estimate. An opt-out page we could not check is marked as unchecked, not assumed live or dead.
Broker records come from the California, Texas, Oregon and Vermont data broker registries, plus parent companies, subsidiaries and government buyers added by research. 93 records appear in at least one published link; most registered brokers have no documented link yet. The README in the zip has the full methods, the record shapes and how to rebuild everything.
Reach
The hubs
Reach is 1 plus the number of records downstream of a broker along published links. It chains separately documented links together; it is not market size, and it does not mean any particular data travelled the whole chain.
| Company | Reach |
|---|---|
| TransUnion | 18 |
| Experian plc | 16 |
| RELX PLC | 14 |
| General Motors LLC | 11 |
| Equifax Inc. | 9 |
| Experian Information Solutions, Inc | 7 |
| LexisNexis Risk Solutions Inc. | 7 |
| TransUnion Risk and Alternative Data Solutions, Inc. | 7 |
| Avast | 6 |
| Babel Street, Inc. | 6 |
| West Publishing Corp | 6 |
| Clearview AI, Inc. | 5 |
CC BY 4.0
Cite, reuse, correct
Reclaim City open data broker dataset, as of 2026-10-24. https://themultiverse.school/research/data-brokers (CC BY 4.0).
Free to share and adapt, including commercially, with attribution (CC BY 4.0). When you cite a specific link, also cite the source document it rests on: the dataset points to the evidence, it is not the evidence.
Found a claim that is wrong, two records that are the same company, or an opt-out route that changed? Open an issue with the record or link ids and, if you can, a document that shows the correction. The dataset is maintained in the Reclaim City repository; this copy is the snapshot at commit 7c0a828c9.