What science mapping adds to a literature review
Science mapping is the relational side of bibliometrics. Where performance analysis tells you how big a field is and who leads it, science mapping shows how the field is organised: which concepts cluster together, which classic works anchor different schools of thought and which groups of researchers collaborate. Zupic and Čater (2015) set out a widely used framework that links each mapping technique to the kind of structure it reveals, and that framework guides how we design projects.
We treat mapping as research design rather than a drawing exercise. Before running anything, we agree with you on the question each map must answer, the unit of analysis and the way time will be sliced. The result is a small set of purposeful figures instead of a gallery of networks that reviewers struggle to connect to the argument.
Conceptual, intellectual and social structure
Conceptual structure
What the field talks about. Usually derived from keyword co-occurrence or terms extracted from titles and abstracts.
Intellectual structure
What the field builds on. Revealed through co-citation of documents, authors or journals in reference lists.
Social structure
Who works with whom. Captured by co-authorship networks at author, institution and country level.
Few projects need all three. A review of an emerging technology may focus on conceptual structure and its evolution, whereas a paper tracing the theoretical roots of a management concept will lean on intellectual structure. Choosing deliberately also keeps word counts under control in journals with tight limits.
Matching techniques to questions
| If your question is | Use | Main output |
|---|---|---|
| Which themes make up the field? | Keyword co-occurrence | Keyword network with thematic clusters |
| Which foundations does it rest on? | Co-citation analysis | Clusters of highly cited classic works |
| Where is the current research front heading? | Bibliographic coupling (Kessler, 1963) | Clusters of recent documents |
| How do groups and countries collaborate? | Co-authorship | Collaboration network |
| Which themes are driving the field and which are fading? | Strategic thematic map | Four-quadrant diagram |
| How have themes changed over time? | Thematic evolution | Flow diagram across periods |
Strategic thematic maps and thematic evolution
The thematic map in Bibliometrix positions each theme on two axes. Centrality measures how strongly a theme connects to the rest of the field; density measures how well developed the theme is internally. The resulting quadrants are commonly read as motor themes, basic themes, niche themes and emerging or declining themes.
When the data is split into periods, a thematic evolution diagram shows which themes inherit from which. Period boundaries strongly influence the picture, so we always state why particular cut points were chosen, for instance a policy change or equal numbers of documents per period, and we can run a sensitivity check with an alternative split.
From research question to finished maps
Design
We agree the questions, the structures to map, the units of analysis and the period boundaries, and record them in a short design note.
Data assembly
Records come from Scopus and Web of Science, and in the Comprehensive package optionally Dimensions, OpenAlex or PubMed. Sources are deduplicated and fields harmonised.
Harmonisation
Keyword variants and cited reference variants are merged so that networks reflect concepts and works rather than spelling noise.
Mapping
Networks are built in VOSviewer; thematic maps, thematic evolution and three-field plots in Bibliometrix. We cross-check that both tools tell a consistent story.
Cluster synthesis and hand-over
A cluster summary table lists members, representative documents and indicators. The Comprehensive package adds a 60-minute walkthrough session.
Cross-checking matters because VOSviewer and Bibliometrix use different clustering and layout methods. When a theme appears clearly in both, you can present it with confidence. When the tools disagree, for example a topic that forms its own cluster in one map but merges with a neighbour in the other, we flag it in the report so you can decide how to handle it in the text rather than discovering the inconsistency during peer review.
Reporting science mapping transparently
Reviewers expect the methods section to justify each technique, state the unit of analysis and thresholds, name the clustering algorithm, explain how centrality and density were computed and describe the period boundaries. Our technical report covers each point and cites the original methodological sources so you can reference them directly. With the Comprehensive package, one round of updates in response to peer review within 60 days is included.
A strong paper also acknowledges the limits of the maps. Database coverage varies across disciplines and languages, indexer-assigned keywords change over time, and recent publications have had little time to accumulate citations, which affects co-citation and coupling results differently. We list the limitations that apply to your dataset in plain terms, so the limitations paragraph you write is specific to your study rather than generic boilerplate.
Division of responsibility
Licensed exports from Scopus or Web of Science are processed only for your project, through your institution's legitimate access where the licence requires it, and are never resold or shared with third parties. Open sources such as OpenAlex are cited according to their own terms.
If you want to understand the mechanics first, see our keyword map workflow guide, check pricing or place an order.