Polish Public Tenders Dataset (BZP + TED) — 2026.Q2

Open dataset of Polish public procurement notices aggregated from official sources: BZP (Biuletyn Zamówień Publicznych, the Polish national procurement bulletin) and TED (Tenders Electronic Daily, the EU-wide procurement database). Coverage: 2024 onward, ~1.4 million tender notices, ~23k buyer profiles, ~82k contractor profiles. Methodology: BZP is ingested via the OCDS (Open Contracting Data Standard) JSON API; TED via its eForms XML feed. TED notices that duplicate a BZP entry are flagged via is_duplicate. Cities are geocoded; provinces normalized to NUTS-2 codes; NIPs (Polish tax IDs) normalized to the canonical 10-digit form. Anonymization: natural-person contractors (CEIDG sole proprietors, PESEL holders) are anonymized via a stable salted SHA-256 hash to comply with Polish/EU data-protection law. Buyers are not anonymized — they are, by law, public bodies. Source code, schema documentation, a getting-started notebook, and the anonymization code (MIT-licensed) are available on GitHub: github.com/atlasprzetargow/polish-tenders-dataset.