BibFusion: A Python package to integrate, deduplicate, and harmonize exported bibliographic records from Scopus and Web of Science for bibliometric analysis

Objective. The study presented BibFusion, a Python software package that harmonizes bibliographic exports from Scopus and Web of Science into a single, traceable, analysis-ready corpus for bibliometric and scientometric research.

Design/Methodology/Approach. BibFusion was capable of ingesting Scopus CSV and WoS TXT files, applying systematic normalization (e.g., ASCII/uppercase standardization of titles and SR keys, affiliation parsing with country extraction), and optionally enriching records via DOI‑based resolution against OpenAlex to recover persistent identifiers (e.g., OpenAlex work IDs, ORCID when available, and OpenAlex author IDs). Cross-database integration employed a DOI-first deduplication cascade with a conservative fallback (title–year–first author) in the event that a DOI is absent. The authors were disambiguated through a canonical PersonID hierarchy (ORCID → OpenAlexAuthorID → normalized name). Citation strings were cleaned and remapped to ensure the preservation of consistent citation links, and journal/Scimago information was consolidated using ISSN/EISSN rules.

Results. In a demonstration on an entrepreneurial marketing query, BibFusion consolidated 436 source records into 253 unique main works and materialized a unified corpus of 8,569 articles. The resulting dataset demonstrated high levels of identifier and geographic completeness, and it provided an analysis-ready citation layer.

Conclusions/Value. BibFusion offers a reusable, auditable integration workflow that has been demonstrated to reduce duplicate inflation and metadata fragmentation. This workflow facilitates the explicit determination of merge decisions and residual uncertainty, thereby ensuring transparency in downstream analyses.