VOSviewer-Assisted Co-Word Analysis
Also known as: keyword co-occurrence analysis via VOSviewer, VOSviewer co-word mapping, keyword network mapping, co-keyword analysis
VOSviewer-assisted co-word analysis is a scientometric pipeline that constructs and visualizes keyword co-occurrence networks from a bibliographic corpus using VOSviewer software. By mapping how often pairs of author-assigned or index keywords appear together in the same publications, the method reveals the intellectual structure of a research field — its dominant themes, emerging topics, and conceptual clusters — producing interactive density and network maps that support systematic interpretation.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use VOSviewer-assisted co-word analysis when the primary research objective is to map the thematic structure, main research clusters, or conceptual evolution of a scientific field from its publication record. It is well suited to large corpora (hundreds to thousands of records) where manual reading is impractical and keyword metadata is consistently populated. The method is appropriate as a standalone bibliometric study or as the science-mapping component of a systematic or scoping review. Do not apply it when keyword fields are sparse or inconsistently filled (common in older literature or some humanities databases), when the corpus is very small (fewer than ~100 records), or when the goal is to synthesize findings rather than map structure — in those cases, qualitative meta-synthesis or meta-ethnography is more appropriate.
Strengths & limitations
- Handles large corpora efficiently — VOSviewer processes thousands of records and thousands of keyword nodes within minutes.
- Provides an intuitive, publication-ready visual map of a field's intellectual structure that supports rapid orientation for newcomers and gap identification for experienced researchers.
- Association strength normalization reduces artefacts caused by high-frequency generic terms dominating the network.
- Freely available open-source software (VOSviewer) with active development support and extensive documentation reduces the technical barrier.
- Complementary to other bibliometric techniques: co-word maps pair naturally with co-citation or bibliographic coupling analyses for a multi-perspective view of a field.
- Results depend heavily on keyword quality: inconsistent author tagging, missing keywords, or absence of index terms degrades the network substantially.
- VOSviewer's community detection has a single resolution parameter that may not capture all relevant granularities; very large or hierarchically structured fields may require iterative parameter tuning.
- Co-word analysis reflects what authors choose to label their work, not necessarily the full conceptual content of the paper, so emerging concepts not yet canonized as keywords may be invisible.
- The method is descriptive — it reveals structure but does not explain causal relationships between research themes or why clusters formed.
Frequently asked
Which keyword field should I use — author keywords or index keywords?
Author keywords (DE in Web of Science) reflect the authors' own framing and are preferred when conceptual mapping is the goal. Index keywords (ID, such as MeSH or KeyWords Plus) provide standardized vocabulary and better coverage when author keywords are sparse. Many studies use both fields combined; if you do, apply thesaurus normalization to merge duplicates between the two sources.
How large does my corpus need to be for a meaningful co-word map?
A practical minimum is around 100 records, but meaningful cluster structure typically emerges from 300 or more records with consistent keyword metadata. Very small corpora produce sparse networks where the minimum occurrence threshold filters out nearly everything. If your corpus is small, consider lowering the threshold to 2, or supplement with a qualitative thematic review of the titles and abstracts instead.
Is VOSviewer-assisted co-word analysis the same as a topic model?
No. Topic modeling (e.g., LDA) derives latent themes from the full text or abstract content of documents using probabilistic methods. Co-word analysis operates on explicit keyword metadata, is deterministic given the network construction settings, and produces a network graph rather than a probability distribution over words. Co-word analysis is faster and more interpretable but depends entirely on keyword quality; topic modeling can work without keyword fields but requires full text and more computational setup.
How do I report a VOSviewer co-word analysis in a journal article?
Report the database(s) searched, the search string, the date of retrieval, the total number of records retrieved and retained, the keyword fields used, the minimum occurrence threshold, any thesaurus file applied, the VOSviewer version, and the resolution parameter used for clustering. Include the network visualization as a figure and provide a supplementary keyword frequency table. This level of transparency allows replication and audit of your mapping decisions.
Can I combine co-word analysis with co-citation analysis in the same study?
Yes, and doing so is increasingly common. Co-word analysis maps the conceptual (thematic) structure of a field based on what researchers write about, while co-citation analysis maps the intellectual base of a field based on what researchers cite. Using both in the same study provides complementary perspectives — conceptual clusters from co-word maps can be compared with citation-based clusters to assess how tightly linked themes are to specific theoretical traditions.
Sources
- van Eck, N. J., & Waltman, L. (2010). Software survey: VOSviewer, a computer program for bibliometric mapping. Scientometrics, 84(2), 523-538. DOI: 10.1007/s11192-009-0146-3 ↗
- Callon, M., Courtial, J. P., Turner, W. A., & Bauin, S. (1983). From translations to problematic networks: An introduction to co-word analysis. Social Science Information, 22(2), 191-235. DOI: 10.1177/053901883022002003 ↗
How to cite this page
ScholarGate. (2026, June 3). VOSviewer-Assisted Co-Word Analysis. ScholarGate. https://scholargate.app/en/scientometrics/vosviewer-assisted-co-word-analysis
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Bibliographic CouplingBibliometrics↔ compare
- Bibliometric AnalysisScientometrics↔ compare
- Co-Citation AnalysisBibliometrics↔ compare
- Science MappingBibliometrics↔ compare
- Scientometric AnalysisScientometrics↔ compare
- Thematic Evolution AnalysisScientometrics↔ compare