Bibliometrix-Assisted Thematic Evolution Analysis

As a practical rule, each time slice should contain at least 30-50 documents to form stable co-occurrence networks. For the overall corpus this typically means 500 or more records spread across at least two time periods. With smaller corpora, themes will be highly sensitive to single documents and the evolution diagram becomes unreliable.

VOSviewer is a standalone GUI application focused on network visualisation, while bibliometrix is an R package that integrates data import, cleaning, network analysis, and strategic diagram construction into a scriptable, reproducible pipeline. bibliometrix provides the thematicMap() and thematicEvolution() functions natively; in VOSviewer, time-sliced maps must be produced manually for each period and compared visually without automated linkage metrics.

Yes. biblioshiny is a Shiny-based graphical interface for bibliometrix that allows users to perform the full thematic evolution workflow through point-and-click menus without writing R code. It is suitable for exploratory work, but scripted R code is preferred when reproducibility and automation are priorities.

Start by defining substantively meaningful periods — for example, before and after a landmark publication, policy change, or technological breakthrough in your field. If no such events are obvious, equal-width windows of 3-5 years are conventional. Check that each slice contains enough documents for stable clustering before finalising the partition.

No. Thematic evolution analysis reveals the structural and topical landscape of a literature — which topics exist, how central they are, and how they shift over time. It does not assess study quality, synthesise evidence on intervention effects, or replace the critical reading required in a qualitative systematic or narrative review. It is most valuable as a complementary component that contextualises and orients a deeper evidence synthesis.