Dataset of 400,000 Prozorro public procurement tender records preserved in synchronization-feed order, with derived award-lot tables
- 1National Board of Revenue, Government of Bangladesh, Rajashwa Bhaban, Plot F-1/A, Agargaon, Sher-e-Bangla Nagar, Dhaka 1207, Bangladesh
Abstract
Public procurement platforms often expose records through synchronization interfaces that order them by last modification rather than by the date of the procurement events they describe. This dataset preserves one such acquisition from Ukraine's Prozorro/OpenProcurement platform. On 13 July 2026, the public tender feed (API version 2.5) was walked from the modification-time offset 1 December 2023 until a ceiling of 400,000 tender identifiers was reached, and each tender was retrieved individually, with no logged failures. The retrieved tenders carry final modification timestamps between 1 and 26 December 2023. The deposit contains the identifiers in the exact order returned, a field-minimized JSON Lines snapshot of the 400,000 records retaining modification time and the buyer, lot, bid, award and award-complaint structure, and three derived tables: a tender-level feed-order table (400,000 rows), an eligible award-lot table (33,148 rows from 31,967 tenders and 8559 buyers) and a fixed November–December 2023 award-lot table (32,800 rows). Identifiers are replaced by consistent pseudonyms; names, contact details, monetary values, procedure types, item classifications, narrative text and documents are excluded. The repository also provides the acquisition script, offline derivation code, a separate re-derivation of the derived tables from the raw files, verification code with regression tests, pinned environments, checksums, a data dictionary and a provenance record listing known gaps. The data support methodological research on how modification-ordered interfaces and record ceilings shape retrieved collections, on stopping rules and cohort definitions, and on how derivation rules determine procedural markers such as single eligible bids and award-level complaint-type objects, as well as teaching reproducible handling of administrative records obtained through application programming interfaces.
Related
- Drug-cost trajectories after national volume-based procurement in Chinese public hospitals: a 31-province interrupted time-series analysis
- Cost estimate for the procurement and preparation of donated corneas for transplantation in an eye bank
- Multi-compartment profiling of donor-derived microRNAs during liver graft procurement: a prospective feasibility study.
- Evidence‑to‑procurement clinical governance in the Emilia‑Romagna TAVI case
- Ethiopia’s first health-sector public–private partnership: a qualitative evaluation of the design, procurement, and early implementation of an integrated diagnostic center
BibTeX
@article{paul2026dataset,
title = {{Dataset of 400,000 Prozorro public procurement tender records preserved in synchronization-feed order, with derived award-lot tables}},
author = {Sushanta Paul},
journal = {Data in Brief},
year = {2026},
month = aug,
doi = {10.1016/j.dib.2026.113365},
url = {https://doi.org/10.1016/j.dib.2026.113365},
note = {CC BY},
}
RIS
TY - JOUR
TI - Dataset of 400,000 Prozorro public procurement tender records preserved in synchronization-feed order, with derived award-lot tables
AU - Sushanta Paul
PY - 2026
DA - 2026/08/01
JO - Data in Brief
DO - 10.1016/j.dib.2026.113365
UR - https://doi.org/10.1016/j.dib.2026.113365
AB - Public procurement platforms often expose records through synchronization interfaces that order them by last modification rather than by the date of the procurement events they describe. This dataset preserves one such acquisition from Ukraine's Prozorro/OpenProcurement platform. On 13 July 2026, the public tender feed (API version 2.5) was walked from the modification-time offset 1 December 2023 until a ceiling of 400,000 tender identifiers was reached, and each tender was retrieved individually, with no logged failures. The retrieved tenders carry final modification timestamps between 1 and 26 December 2023. The deposit contains the identifiers in the exact order returned, a field-minimized JSON Lines snapshot of the 400,000 records retaining modification time and the buyer, lot, bid, award and award-complaint structure, and three derived tables: a tender-level feed-order table (400,000 rows), an eligible award-lot table (33,148 rows from 31,967 tenders and 8559 buyers) and a fixed November–December 2023 award-lot table (32,800 rows). Identifiers are replaced by consistent pseudonyms; names, contact details, monetary values, procedure types, item classifications, narrative text and documents are excluded. The repository also provides the acquisition script, offline derivation code, a separate re-derivation of the derived tables from the raw files, verification code with regression tests, pinned environments, checksums, a data dictionary and a provenance record listing known gaps. The data support methodological research on how modification-ordered interfaces and record ceilings shape retrieved collections, on stopping rules and cohort definitions, and on how derivation rules determine procedural markers such as single eligible bids and award-level complaint-type objects, as well as teaching reproducible handling of administrative records obtained through application programming interfaces.
SN - 2352-3409
ER -