Data sources
Last updated: 2026-10-23.
Every Arcadium Data dataset comes from sources that anyone can read without an account: official open-data services under their published licences, and public pages of businesses. This page lists them per dataset. How the events are built from them is on Methodology.
Public tenders and awards
Calls for tenders, contract awards and their winners, buyers, values and CPV codes.
| Source | How we read it | Licence or basis |
|---|---|---|
| EU Tenders Electronic Daily (TED), Publications Office of the EU | Official TED Search API, anonymous, within the published fair-use limits | Reuse under Commission Decision 2011/833/EU, with attribution "© European Union" |
| UK Find a Tender service, Cabinet Office | Official OCDS API, anonymous | Open Government Licence v3.0 |
| France: BOAMP open data, DILA | Official open-data API (Opendatasoft), anonymous | Licence Ouverte / Open Licence 2.0 (Etalab) |
| Germany: oeffentlichevergabe.de (Datenservice Öffentlicher Einkauf) | Official daily notice exports, anonymous | CC0 1.0, as stated in every export |
| Poland: Biuletyn Zamówień Publicznych (e-Zamówienia) | The public notice board's own anonymous endpoint | Official notices are not protected by copyright (Polish Copyright Act, art. 4); reused as public sector information |
| Bulgaria: Public Procurement Agency (CAIS EOP) on data.egov.bg | National open-data portal API, anonymous | CC0 1.0, as stated in every package |
Jobs feed, hiring signals and posted pay
Open jobs per employer (title, location, department, posted pay range) and the counts and changes derived from them.
| Source | How we read it | Licence or basis |
|---|---|---|
| Employer career sites and their public job boards on applicant tracking systems (for example Greenhouse, Lever, Ashby, SmartRecruiters, Workday) | The public job-board pages and the endpoints those pages load for every visitor, no login | Public job advertisements published by the employer; read at low request rates |
| Sweden: JobTech (Arbetsförmedlingen) JobStream | Official open API, anonymous | CC0 1.0, as stated in the API documents |
| Public job listing sites | Public listing pages, no login | Public job advertisements; employer-level fields only |
| Common Crawl web archive (hiring history panel) | Public crawl indexes and archived captures of public job boards | Common Crawl terms of use |
Shopify stores and installed apps
Public storefront facts: country, currency, catalogue size, apps and tools visible on the store.
| Source | How we read it | Licence or basis |
|---|---|---|
| Public storefront pages of online stores | The pages and the endpoints those pages load for every visitor, no login | Publicly shown business information; stores whose name reads like a private person are left out |
Review panels
Ratings, review counts and review velocity per company or product.
| Source | How we read it | Licence or basis |
|---|---|---|
| Public review pages on review platforms | Public pages, no login, no account | We publish counts, ratings and trends; reviewer names are never stored or published |
Hotel demand curve
Forward nightly rates of a fixed hotel panel per city and the demand index derived from them.
| Source | How we read it | Licence or basis |
|---|---|---|
| Public hotel rate listings | Public search result pages, no login, no booking | Publicly shown prices of businesses |
Ad radar
Counts of active, new and stopped ads per brand and platform.
| Source | How we read it | Licence or basis |
|---|---|---|
| Public ad transparency libraries of the large online platforms | Public library search, no login | Ad repositories the platforms publish under the EU Digital Services Act (art. 39) |
Company Pulse
Company Pulse adds no source of its own: its events are derived from the datasets above (jobs, posted pay, tools named in job ads, store scans, review panels and contract awards) and joined on the company's website domain.
How we collect
- Only data that is public without a login. We do not create accounts, use anyone's credentials or tokens, or get around paywalls.
- Official open-data services are used through their documented APIs and within their published rate limits. Their licences travel with the data: attribution lines are in each dataset's README.
- For public web pages we review each site's robots.txt and terms before we start and record what they say. We keep request rates low, read only the pages and fields we need, and slow down or stop when a site signals it.
- The data is about organisations. Names, emails and phone numbers of people are removed before anything is stored, and free text is cleaned again on the way out.
Removal requests
If you run a site we read, or a record is about you or your business, write to hello@arcadiumdata.io with the source or the record (a domain or a URL is enough). We stop collecting it and remove it from the files, the API and the data pages within 24 hours, keep it out of later exports, and reply once it is done. The same policy is on the legal page.