MLKN.lab · Data
Data
The data sources, methodologies, datasets, and tools behind MLKN.lab’s knowledge networks.
Overview
The data sources, methodologies, datasets, and tools
The empirical foundation of MLKN.lab relies on a curated, high-fidelity knowledge repository aggregated from massive-scale, open-access and proprietary bibliometric corpora. By harvesting metadata from over 200 million scholarly records via OpenAlex, and enriching them with citation depth from Scopus alongside specialized taxonomies like MeSH and the IEEE Thesaurus, the laboratory maintains a comprehensive cross-domain registry.
Through automated pipelines executing rigorous deduplication, disambiguation, and standardization against international benchmarks (OECD Frascati and UNESCO metrics), raw bibliographic streams are cross-compiled into multi-layered JSON and CSV graphs. This infrastructure exposes a network architecture of 25 macro-disciplines, 235 specialized subdisciplines, and more than 320,000 structural connections, providing researchers with an open-access, analytically validated substrate for advanced knowledge graph mining and computational epistemology.
OpenAlex
Data Sources Introduction
OpenAlex: The Backbone of MLKN.lab’s Data
MLKN.lab relies on OpenAlex, a free, open catalog of scholarly papers, to build its polyhierarchical knowledge hypergraph. OpenAlex provides:
- Comprehensive Metadata: Titles, authors, abstracts, citations, and more for over 200M works.
- Interconnected Data: Relationships between papers, authors, institutions, and concepts.
- Open Access: Free to use, with no restrictions on commercial or non-commercial use.
- Real-Time Updates: Continuously updated to reflect the latest research.
By using OpenAlex, MLKN.lab aligns with the global movement toward open science, ensuring that our knowledge hypergraph is transparent, reproducible, and sovereign.
Data Sources
Where Our Knowledge Comes From
OpenAlex
The primary data source for MLKN.lab. OpenAlex is a free, open catalog of scholarly papers, authors, venues, and institutions, with over 200M works.
Scopus
A comprehensive abstract and citation database of peer-reviewed literature, used to supplement OpenAlex data for thematic and conceptual mappings.
MeSH (Medical Subject Headings)
A controlled vocabulary thesaurus produced by the U.S. National Library of Medicine. Used for biomedical and health science classifications.
IEEE Thesaurus
A controlled vocabulary for indexing and retrieving engineering, computing, and technology literature. Used for technical and applied science classifications.
Methodology
How We Process and Structure Data
Data Cleaning
Raw data from OpenAlex and other sources is cleaned, deduplicated, and standardized to ensure consistency across disciplines.
Classification
Disciplines and subdisciplines are classified using OECD Frascati Manual and UNESCO Fields of Science standards.
Network Construction
We map connections between disciplines using co-occurrence analysis and citation networks to reveal interdisciplinary links.
Validation
Our hierarchy is validated through network metrics (e.g., centrality, modularity) and expert review to ensure accuracy.
Datasets
Access Our Data
MLKN.lab Hierarchy Master File (All Layers, All Details)
The complete hierarchy dataset for MLKN.lab, containing all layers and details of the polyhierarchical knowledge network. This file is hosted on Zenodo for reliability and long-term preservation.
Size: 92.3 MB | Format: CSV | DOI: 10.5281/zenodo.21363227 | Version: v1.0
Download Dataset View on Zenodo README
Citation: Papin, F. (2026). MLKN.lab: A Polyhierarchical Hypergraph of Scientific Knowledge for Metascience, Computational Epistemology, and Network Analysis [Dataset]. Zenodo. https://doi.org/10.5281/zenodo.21363227
Full Hierarchy Datasets
Complete datasets for the interdisciplinary knowledge network, including all 25 disciplines, 235 subdisciplines, and 320K+ connections.
Network Data (JSON)
Structured JSON files for nodes and edges, optimized for visualization and analysis.
Network Data (CSV)
Tabular CSV files for nodes, edges, and mappings, compatible with most data analysis tools.
Knowledge Network Data (Mirror)
Duplicate copies of the above datasets, stored in knowledge_network/data/ for direct use in visualizations.
Tools
Software and Scripts for Working with Our Data
Network Analysis Scripts
Python scripts for analyzing the knowledge networks, including centrality metrics, community detection, and visualization.
Data Processing Pipeline
A modular pipeline for cleaning, classifying, and structuring raw data from OpenAlex and other sources.
Citation
How to Cite MLKN.lab
If you use MLKN.lab or its datasets in your research, please cite the following:
For the Software
Papin, F. (2026). MLKN.lab: A Polyhierarchical Framework for Modeling Scientific Knowledge [Software]. GitHub. https://github.com/FrancoisPapin/MLKN-lab
For the Master File Dataset
Papin, F. (2026). MLKN.lab: A Polyhierarchical Hypergraph of Scientific Knowledge for Metascience, Computational Epistemology, and Network Analysis [Dataset]. Zenodo. https://doi.org/10.5281/zenodo.21363227
Collaborate with Us
Interested in using our data, tools, or methodology? Let’s work together.
We’re open to collaborations in meta-science, network analysis, knowledge graphs, and interdisciplinary research.