Bibliometric analysis has become an important method for researchers who want to understand the structure, development, and intellectual landscape of a research field. It is widely used in systematic reviews, literature mapping, science mapping, research evaluation, and trend analysis.
Two of the most popular tools for bibliometric research are VOSviewer and Biblioshiny. VOSviewer is especially useful for creating and visualising bibliometric networks, while Biblioshiny provides a user-friendly interface for the Bibliometrix package in R and supports a broader range of performance and science-mapping analyses.
Used together, these tools can help researchers identify influential authors, leading journals, major countries, highly cited papers, research themes, collaboration networks, keyword clusters, and emerging topics.
What Is Bibliometric Analysis?
Bibliometric analysis is the quantitative study of academic literature using publication and citation data.
It can be used to answer questions such as:
Which authors are most productive?
Which papers are most highly cited?
Which journals publish the most research on a topic?
Which countries or institutions collaborate most frequently?
What are the major research themes?
How has the field evolved over time?
Which keywords are becoming more important?
Bibliometric analysis is different from a traditional literature review because it focuses on large-scale patterns in publication metadata.
It does not replace critical reading, but it can help researchers understand the structure of a field before conducting deeper qualitative interpretation.
Step 1: Define the Research Question
A bibliometric study should begin with a clear objective.
Examples include:
mapping research trends in sustainable transportation,
identifying major themes in artificial intelligence research,
analysing international collaboration in climate change studies,
examining the development of transit-oriented development literature,
or identifying emerging research topics in urban planning.
The research question determines the search strategy, database, time period, document types, and analysis techniques.
A vague objective often leads to weak bibliometric results.
Step 2: Select the Database
The next step is to choose an appropriate bibliographic database.
Common options include:
Scopus,
Web of Science,
Dimensions,
PubMed,
and other subject-specific databases.
Scopus and Web of Science are especially common in bibliometric studies because they provide structured metadata and citation information.
Researchers should avoid combining datasets from multiple databases without a clear deduplication and harmonisation strategy.
Different databases may index different journals and record citations differently.
The chosen database should therefore be clearly stated in the methodology.
Step 3: Develop a Search Strategy
The search strategy should be transparent and reproducible.
Researchers should identify:
keywords,
synonyms,
Boolean operators,
subject terms,
date range,
language,
document type,
and subject area.
For example, a study on transit-oriented development might use:
“transit-oriented development” OR “transit oriented development” OR “TOD”
The researcher may then combine these terms with concepts such as travel behaviour, land use, accessibility, or sustainability.
Search strings should be tested carefully before downloading data.
A search that is too narrow may miss important papers, while a search that is too broad may retrieve irrelevant records.
Step 4: Apply Inclusion and Exclusion Criteria
The dataset should be cleaned before analysis.
Researchers may exclude:
editorials,
notes,
conference abstracts,
unrelated subject areas,
non-English papers,
duplicate records,
or publications outside the selected time period.
The criteria should be decided in advance and reported clearly.
If the bibliometric study is combined with a systematic literature review, researchers may also use PRISMA-style screening to document record selection.
Step 5: Export Bibliographic Data
Once the final dataset is identified, export the records in a format compatible with VOSviewer or Biblioshiny.
Useful metadata may include:
author names,
titles,
abstracts,
keywords,
affiliations,
publication year,
journal title,
references,
citation counts,
and DOI.
The export format depends on the database.
Researchers should make sure that citation and reference information is included if they plan to conduct co-citation or bibliographic coupling analysis.
Step 6: Clean the Data
Data cleaning is one of the most important stages.
Common problems include:
duplicate author names,
inconsistent institution names,
keyword variations,
abbreviations,
spelling differences,
singular and plural terms,
and inconsistent country names.
For example:
“AI,” “artificial intelligence,” and “Artificial Intelligence” may represent the same concept.
Similarly:
“USA,” “United States,” and “United States of America” should usually be standardised.
Poor data cleaning can produce misleading clusters and fragmented networks.
Step 7: Use VOSviewer for Network Mapping
VOSviewer is particularly effective for visualising bibliometric relationships.
Researchers can create maps based on:
co-authorship,
co-occurrence,
citation,
bibliographic coupling,
and co-citation.
Each type of analysis answers a different question.
Co-Authorship Analysis
Co-authorship analysis examines collaboration among:
authors,
institutions,
or countries.
It can reveal major research networks and identify highly collaborative scholars.
In a country-level analysis, larger nodes generally represent greater publication activity, while links represent collaboration.
This can help researchers understand the international structure of a field.
Keyword Co-Occurrence Analysis
Keyword co-occurrence is one of the most widely used VOSviewer techniques.
It shows which keywords frequently appear together in the same publications.
The resulting network usually contains clusters.
For example, a study on urban transport may produce clusters related to:
accessibility,
public transport,
sustainability,
travel behaviour,
land use,
and mobility.
The clusters can help researchers identify major thematic areas.
However, researchers should not treat automatically generated clusters as final themes without interpretation.
Citation Analysis
Citation analysis identifies influential publications, authors, journals, or institutions based on citation counts.
This can help answer questions such as:
Which article has had the greatest impact?
Which journals are most influential?
Which authors are most frequently cited?
However, citation counts should be interpreted carefully.
Older publications generally have had more time to accumulate citations, and citation practices vary across disciplines.
Co-Citation Analysis
Co-citation occurs when two documents, authors, or journals are cited together by later publications.
Co-citation analysis can help identify the intellectual foundations of a research field.
If two authors are frequently cited together, they may represent related theoretical traditions or research schools.
This type of analysis is useful for understanding the knowledge base underlying a field.
Bibliographic Coupling
Bibliographic coupling examines whether two publications cite the same references.
If two papers share many references, they may be intellectually related.
This method is especially useful for identifying contemporary research clusters because it focuses on similarities in reference lists.
Step 8: Interpret VOSviewer Maps Carefully
VOSviewer visualisations typically use:
nodes,
links,
clusters,
and distances.
A larger node generally indicates greater weight.
A stronger link indicates a stronger relationship.
Colours often represent clusters.
Distance can indicate relatedness.
However, researchers should avoid making overly strong conclusions from visual appearance alone.
The interpretation should be supported by quantitative indicators and close reading of representative publications.
Step 9: Use Biblioshiny for Descriptive Analysis
Biblioshiny is a graphical interface for Bibliometrix in R.
It is especially useful for researchers who want advanced bibliometric analysis without writing extensive code.
Biblioshiny can generate:
annual publication trends,
most productive authors,
most relevant journals,
most cited documents,
country productivity,
institutional productivity,
keyword frequency,
thematic maps,
conceptual structures,
collaboration networks,
and source impact.
This makes it highly useful for preparing tables and figures for research papers.
Step 10: Analyse Annual Scientific Production
One of the first analyses in Biblioshiny is publication growth over time.
Researchers can examine:
number of publications per year,
annual growth rate,
periods of rapid expansion,
and historical development.
A sudden increase in publications may indicate that a topic is becoming more important.
However, publication growth should be interpreted in context.
An increase may also reflect broader growth in academic publishing.
Step 11: Identify Leading Authors, Journals and Countries
Biblioshiny allows researchers to identify the most productive:
authors,
journals,
institutions,
and countries.
This helps describe the research landscape.
Researchers can also examine citation impact, which provides a different perspective from publication volume.
An author with fewer papers may be more influential if those papers receive many citations.
Productivity and impact should therefore be reported separately.
Step 12: Conduct Thematic Analysis
One of Biblioshiny’s useful features is thematic mapping.
Themes may be classified according to dimensions such as centrality and development.
Researchers can identify:
motor themes,
emerging themes,
basic themes,
and specialised themes.
This can help explain how the intellectual structure of a field is evolving.
Thematic evolution analysis can also show how research topics change over different time periods.
Step 13: Combine VOSviewer and Biblioshiny
The two tools are most powerful when used together.
VOSviewer is excellent for:
visual network maps,
clustering,
co-authorship,
keyword co-occurrence,
and citation relationships.
Biblioshiny is strong for:
descriptive statistics,
productivity analysis,
thematic mapping,
trend analysis,
and broader bibliometric summaries.
A robust study may use Biblioshiny to describe the dataset and VOSviewer to visualise networks.
Step 14: Validate the Results
Bibliometric analysis should not rely entirely on software-generated output.
Researchers should read representative articles from major clusters and verify whether the interpretation makes conceptual sense.
For example, if a keyword cluster is labelled “sustainable mobility,” the papers in that cluster should actually reflect that theme.
Software can group terms statistically, but the researcher must assign meaning.
Step 15: Report the Methodology Transparently
A good bibliometric paper should report:
database used,
search date,
search string,
time period,
inclusion criteria,
document types,
number of records,
data-cleaning procedures,
software version,
VOSviewer settings,
Biblioshiny settings,
counting method,
thresholds,
and analysis techniques.
This information improves reproducibility.
Researchers should also explain whether full counting or fractional counting was used where relevant.
Common Mistakes to Avoid
Researchers should avoid several common problems.
These include:
using an unclear search strategy,
failing to clean keywords,
mixing databases without proper deduplication,
interpreting clusters mechanically,
reporting only colourful maps without analytical discussion,
ignoring citation-age bias,
and using software output without explaining the research significance.
A bibliometric study should tell a meaningful story about the field, not simply display graphs.
Conclusion
VOSviewer and Biblioshiny are powerful tools for conducting bibliometric analysis.
VOSviewer is particularly useful for visualising collaboration, keyword, citation, co-citation, and bibliographic coupling networks. Biblioshiny provides a broader analytical environment for examining publication trends, influential authors, journals, countries, conceptual structures, and thematic evolution.
For researchers, the most effective workflow is to begin with a clear research question, select an appropriate database, develop a reproducible search strategy, clean the dataset carefully, and then use both tools according to their strengths.
The central principle is that bibliometric software supports analysis, but interpretation remains the responsibility of the researcher.
For researchers and academic support organisations such as EduPub, combining VOSviewer and Biblioshiny can provide a systematic and visually powerful way to explore large bodies of scholarly literature, identify research trends, map intellectual structures, and discover emerging opportunities for future research.
