VOSviewer-Assisted Thematic Evolution Analysis
Also known as: VOSviewer thematic mapping, keyword co-occurrence thematic evolution, science mapping thematic evolution, VOSviewer longitudinal thematic analysis
VOSviewer-assisted thematic evolution analysis is a scientometric pipeline that uses the VOSviewer software to build keyword co-occurrence networks across chronological time slices of a bibliographic dataset, revealing how research themes emerge, converge, fragment, or disappear over time within a scientific field. By coupling VOSviewer's density-based clustering with period-by-period comparison, researchers obtain a visual and quantitative account of a field's intellectual trajectory.
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 this approach when the research goal is to trace how the intellectual structure of a scientific field has changed over time using bibliographic data. It suits researchers conducting systematic or scoping reviews who want to go beyond frequency counts and reveal dynamic thematic patterns. Appropriate for fields with at least several hundred indexed records spanning a minimum of ten years. Not suitable when the publication base is too small (fewer than 200 records), when keyword indexing is inconsistent across the corpus, or when the research question requires qualitative reading of primary study content rather than bibliometric mapping. It complements but does not replace a full systematic review of study findings.
Strengths & limitations
- VOSviewer is freely available, actively maintained, and widely validated in the scientometric community, ensuring methodological transparency and reproducibility.
- Reveals the longitudinal dynamics of a field's thematic structure that static bibliometric analyses cannot capture.
- Produces high-quality, exportable network visualizations suitable for academic publication.
- Handles large bibliographic datasets (tens of thousands of records) efficiently, making it scalable to comprehensive literature searches.
- Quantitative grounding through co-occurrence frequencies provides an objective basis for thematic classification.
- Dependent on keyword quality: poorly assigned or inconsistent author keywords produce misleading clusters; index keywords (MeSH, WoS categories) are more reliable but narrower.
- Cluster labels are assigned by the researcher through interpretation of the most frequent terms — this introduces a degree of subjectivity and requires domain expertise.
- Thematic continuity across time slices must be inferred by the researcher, as VOSviewer does not automatically link clusters across periods.
- Does not capture the content or quality of individual studies — high-frequency themes may reflect trendy but methodologically weak research streams.
- Limited to topics with substantial indexed bibliographic coverage; emerging or interdisciplinary fields may be underrepresented in major databases.
Frequently asked
How is this different from a simple keyword frequency trend analysis?
Keyword frequency trend analysis counts how often individual terms appear per year, treating each keyword independently. VOSviewer-assisted thematic evolution analysis maps the co-occurrence relationships between keywords, revealing thematic clusters — groups of concepts that travel together in the literature — and then tracks how those clusters shift across time. The network perspective captures thematic structure, not just individual term prevalence.
How many time periods should I divide the dataset into?
There is no universal rule, but three to five periods is common. Periods should be long enough to generate sufficient co-occurrence data per slice (aim for at least 100–150 records per period) while being short enough to detect meaningful change. Five-year windows work well for mature fields with thousands of records; decade windows suit smaller or older corpora. Always report your rationale.
Can I use VOSviewer with Scopus data instead of Web of Science?
Yes. VOSviewer accepts RIS and CSV exports from both Scopus and Web of Science, as well as PubMed. The choice of database affects coverage: Scopus tends to cover more social science and non-English journals, while Web of Science has deeper historical coverage for natural sciences. Using both and cross-checking keyword alignment is best practice for comprehensive reviews.
Do I need to report inter-rater reliability for cluster labeling?
VOSviewer assigns keywords to clusters algorithmically, so the clustering itself does not require inter-rater reliability testing. However, the labels you assign to clusters are interpretive, and many journals expect either a second rater to independently label clusters or a transparent justification of each label based on the top-frequency keywords. Check target journal expectations and report your labeling process explicitly.
Is there a minimum dataset size for this method?
VOSviewer can technically process very small datasets, but thematic evolution analysis requires sufficient records per time slice to form meaningful co-occurrence networks. As a practical minimum, aim for at least 200 records in total and at least 50–100 records per time slice. Below this threshold, few keyword pairs will meet any reasonable co-occurrence threshold, and resulting clusters will be unstable and uninterpretable.
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 ↗
- Cobo, M. J., López-Herrera, A. G., Herrera-Viedma, E., & Herrera, F. (2011). Science mapping software tools: Review, analysis, and cooperative study among tools. Journal of the American Society for Information Science and Technology, 62(7), 1382–1402. DOI: 10.1002/asi.21525 ↗
How to cite this page
ScholarGate. (2026, June 3). VOSviewer-Assisted Thematic Evolution Analysis. ScholarGate. https://scholargate.app/en/scientometrics/vosviewer-assisted-thematic-evolution-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.
- Bibliometric AnalysisScientometrics↔ compare
- Co-Citation AnalysisBibliometrics↔ compare
- Co-word AnalysisScientometrics↔ compare
- Science MappingBibliometrics↔ compare
- Scientometric AnalysisScientometrics↔ compare
- Thematic Evolution AnalysisScientometrics↔ compare