How to export data from Scopus, Web of Science and PubMed for bibliometric analysis

The export settings that keep every field your analysis needs, for each major database, and how to combine the files into one de-duplicated data set.

Which fields you need

Different analyses need different fields. Export the full record whenever you can.

AnalysisFields needed
Annual production, sources, authorsAuthors, title, source, year
Citation indicators (h-index, most cited)Times cited
Keyword analyses and thematic mapsAuthor keywords, Keywords Plus / index keywords, abstract
Countries, institutions, collaborationAffiliations, corresponding author address
Co-citation, coupling, historiograph, RPYS, local citationsCited references

Scopus

  1. Run your search and tick All to select every result.
  2. Click Export and choose CSV (BibTeX and RIS also work).
  3. Under the fields to export, select all information: citation information, bibliographical information, abstract and keywords, funding details, and References.
  4. Click Export. At the time of writing, Scopus allows up to 20,000 records per CSV, RIS or BibTeX export; for larger result sets, split the search by year and export each part.

Web of Science

  1. Search the Web of Science Core Collection (other databases on the platform export fewer fields).
  2. Click Export and choose Plain text file. Tab delimited, BibTeX and Excel also work.
  3. Set the record content to Full Record and Cited References.
  4. Export records 1–1,000, then 1,001–2,000, and so on, because each file is limited to 1,000 records. Drop all the files into the app at once.

PubMed

  1. Run your search and click Save.
  2. Choose All results (up to 10,000 records) and the format PubMed. This produces a MEDLINE-tagged .txt file with MeSH terms, author keywords and affiliations. The CSV format also works but contains fewer fields.

PubMed records have no citation counts and no reference lists, so citation indicators, co-citation and coupling are not available from PubMed alone. A common approach is to search PubMed for coverage and Web of Science or Scopus for citation data, then merge.

Dimensions

From the results list, choose Save / Export and either the CSV or the bibliographic mapping export (which includes cited references where available). The first "About the data" line of the file is skipped automatically. Export limits depend on your account type.

Lens.org

In scholarly search, click Export and choose CSV. Lens exports references as Lens IDs, which supports local citations and bibliographic coupling but not reference labels or reference publication years.

OpenAlex

On openalex.org, filter your works and download them as CSV. Authors, institutions, countries, keywords, topics, citation counts and referenced works (as OpenAlex IDs) are read when present.

Google Scholar and other sources

Google Scholar has no bulk export, but Publish or Perish can save Google Scholar results as CSV or RIS. Zotero, Mendeley, EndNote, IEEE Xplore, EBSCO and ProQuest can all export RIS or BibTeX. These sources usually lack cited references, so use them mainly for performance and keyword analyses.

Merging databases

Drop every export file into Bibliometrics Studio together. Records are merged, and a record found in two databases is kept once: matching uses the DOI first, then the normalised title with the publication year. The fields of the richer record (for example Web of Science cited references) are kept and gaps are filled from the other. The import page reports how many duplicates were removed, and the PRISMA page uses these counts.

Mixing reference formats from different databases can split the same cited work into two labels in co-citation analysis. If co-citation is central to your study, run it on a single database.

Related guides