LOPDF: a framework for extracting and producing open data of scientific documents for smart digital libraries

Area of Science:

• Bibliometrics
• Data Science
• Information Science

Background:

• Scientific publications contain valuable metadata crucial for collaboration and discovery.
• Unstructured metadata hinders smart querying and analysis of research data.
• Processing publication data is complex, time-consuming, and resource-intensive.

Purpose of the Study:

• To address the challenge of utilizing unstructured scientific publication metadata.
• To develop a generic framework for transforming publication data into machine-understandable formats.
• To enable advanced analysis of research impact and behavior.

Main Methods:

• Developed the Linked Open Publications Data Framework (LOPDF).
• Crawls, processes, and extracts data from diverse publisher sources.
• Transforms textual publication data into semantically enriched RDF datasets (Linked Open Data).

Main Results:

• Generated machine-understandable Linked Open Data (LOD) from scientific publications.
• Resulting datasets support smart queries using the SPARQL protocol.
• Quantitative and qualitative analyses demonstrate the utility for computing research behavior and impact.

Conclusions:

• The LOPDF framework facilitates the creation and processing of open data for scientific publications.
• Enables sophisticated analysis of research trends, impact, and collaboration patterns.
• Supports data-driven insights into the evolving knowledge society.