EuroProcure-10: A Large-Scale European Public Procurement Dataset from TED (2016–2025)

EuroProcure-10 is a large-scale machine-learning-ready dataset constructed from Tenders Electronic Daily (TED), the official public procurement portal of the Publications Office of the European Union. The dataset contains 1,500,356 procurement notices published between 2016 and 2025 and provides 51 structured attributes covering procurement characteristics, financial information, geographic metadata, procedural details, contract information, and enriched analytical features. The dataset was created through a multi-stage data engineering pipeline involving TED archive acquisition, XML parsing, schema harmonization, data cleaning, normalization, CPV-based domain classification, and enrichment with additional procurement intelligence attributes. EuroProcure-10 covers seven strategic procurement domains: Healthcare Digital Technology Construction and Infrastructure Automotive Green Energy Defence and Security Aerospace Key enriched attributes include: Procurement domain classification based on CPV codes AI-generated semantic tags and procurement keywords Contract value consistency indicators Award-to-estimate ratio analysis EU funding indicators Competition and procedural characteristics The dataset supports research in machine learning, public procurement analytics, economic analysis, transparency studies, anomaly detection, procurement intelligence, and public-sector AI applications. Raw data were sourced from TED (Tenders Electronic Daily), Publications Office of the European Union, and used under the Creative Commons Attribution 4.0 International (CC BY 4.0) licence. EuroProcure-10 represents an enriched and processed version of the original procurement notices. The European Union, TED, and the Publications Office of the European Union do not endorse this dataset, its derived attributes, analyses, conclusions, or associated products. Citation:10.5281/zenodo.20690042