Published September 16, 2026 | Version v1.1.0

smarqart-hash/ted-procurement-data-quality: v1.1.0 — Revised manuscript replication package

Authors/Creators

Description

Replication package for the revised manuscript "Silent Placeholder Contamination in EU Procurement Data: A Field-Level Audit of TED Contract Values" (Data & Policy, minor revision).

What is new since v1.0.0

  • Analysis scripts (scripts/analysis/): value-free ranking of the top-50 placeholder records, stratified permutation null, composite specification variants, CRI proxy convergence with Shapley attribution, OpenTender.eu benchmark, core rank stability, top-50 indicator breakdown, random and matched removal recomputation, buyer/supplier placeholder analyses.
  • Scoring CLI (scripts/scoring/compute_scores.py): --config (unfiltered v1.1 scores), --exclude-placeholders, --entity-mapping, --output. Entity clustering is now order-independent.
  • Frozen score artifacts in data/processed/: scoring v1.2, unfiltered v1.1 and the recomputation without placeholder records, plus contact-redacted entity mappings, all with SHA-256 checksums. These files are the basis of every figure in the manuscript.
  • README: key findings aligned with the manuscript, replication path (frozen artifacts) and recomputation path, data links; 692 tests.

Manuscript, response letters and review material are not part of this repository.

Files

smarqart-hash/ted-procurement-data-quality-v1.1.0.zip

Files (14.6 MB)

Name Size
md5:6beb98cd84dff3eea5e45b9477ab1152
14.6 MB Preview Download

Additional details