e-Procurement Dataset Based on Indonesian National e-Katalog Services v6.0

This dataset delivers a systematized table of 595,497 products and 8,094 supplier companies across 2,494 categories, providing filterable and query able data to support descriptive statistical analysis and AI-based applications for the Indonesian public procurement phase. The inclusion of an explicit rating dataset, which has been re-scaled, allows researchers in Machine Learning (ML) and Natural Language Processing (NLP) to train and test new models for supervised learning problems, specifically for predicting procurement preferences and improving recommendation accuracy. Researchers and developers can utilize this data to build CRS models that assist government procurement officers (PPK) in navigating thousands of product options through interactive dialogue, tackling the critical blockage of information scarcity in the public sector. Beyond raw data, the provision of the dataset in Ontology Web Language (OWL) format serves as a robust knowledge base. This allows for semantic reasoning and helps address the "cold-start" problem in recommender systems by establishing logical relationships between procurement entities. By offering the dataset in three distinct formats, which is CSV, JSON, and OWL, the data ensures high interoperability. The CSV and JSON formats facilitate quick integration into popular data science pipelines (such as Pandas or Scikit-Learn), while the OWL format supports the development of knowledge-based AI systems. This data enables researchers, construction analysts, and bidders to identify trends in national procurement, perform price benchmarking, and assess the transition performance of the e-Katalog 6.0 system, providing insights into factors that affect a project's success at its early stages.