DABYTE OPEN DATA
Contribute a dataset: your numbers, your name, open to everyone
Answer. If your company holds primary data nobody else has — the discounts you actually grant, telemetry from your own product, results of a benchmark you ran — DABYTE will normalise it and publish it at a permanent address under CC BY 4.0, with your name in the citation line. It stays free for everyone, including your competitors. What you get is not an advertisement next to the data: it is your name inside the fact that answer engines cite. 0 datasets published so far.
Format below · validate before sending ·
schema.json
Why a company pays to give data away
Because an answer engine cites data, not advertising. When somebody asks an assistant a question your data answers, the assistant reaches for the public dataset on that subject — and if the only one is yours, your name travels with the answer for as long as the data stands. Buying a placement puts you beside the answer. Contributing data puts you in it.
The unusual part, and the part that makes it work: we publish it free. Data behind a paywall is not cited, so a dataset only pays back its contributor if anyone can take it. That is also why we insist on CC BY — the licence is the distribution.
What we accept, and what we refuse on sight
Accepted: facts about your own activity that you can stand behind — prices actually paid or offered, discount bands you actually grant, anonymised usage telemetry, results of a reproducible benchmark you ran, incident and uptime statistics for your own infrastructure, verifiable regulatory facts about yourself.
Refused, without exception: anything evaluative about you or anyone else; comparisons naming competitors; a ranking, shortlist or award; marketing copy in any field; a dataset whose collection method you will not state; and any request, in any form, to affect a score in the visibility index. The validator rejects evaluative wording automatically — one «best-in-class» in a cell and the file comes back.
How to send it — four steps, no meeting required
- Shape the file. One JSON envelope:
dataset,contributor,method,rows, plus a licence acknowledgement and a completeness statement. Every row must have the same fields. The full schema with a worked example is at/contribute/schema.json. - Validate it yourself. The same validator we run is public, so nothing is a surprise: it checks required fields, dates, row homogeneity, that your sample size is consistent with the rows, and that no evaluative wording slipped in.
- Send one file. To [email protected], subject «Dataset contribution». No portal, no account.
- Check it landed. Within two working days it is either published or
returned with the exact list of validator errors. When published you get a permanent
/data/<slug>/address, and you can verify placement yourself with one call to/api/datasets.json— your legal name, your domain and the canonical URL are in the record.
What you get, stated precisely
| You get | Where to check it |
|---|---|
| A permanent page | /data/<slug>/, rebuilt but never moved |
| Your name in the citation | the cite line on the page and in every machine copy |
| Machine copies | /api/datasets/<slug>.json, CSV, markdown mirror |
| Presence in the index of datasets | /api/datasets.json, sitemap, Atom feed |
| Submission to answer-engine crawlers | IndexNow ping on publication |
| A link to your site | on the dataset page, rel="nofollow" — attribution, not a ranking favour |
What you explicitly do not get, at any price: a better score, a place in the index, a comparison in your favour, or a say in how anything else on this site is worded.
Published contributions
- No contributed datasets yet. This page is the intake, not a showcase — the first one publishes here the day it passes validation.