Peer-reviewed Water Security, Resilience, and Adaptation Literature for a Bibliometrics and Topic Modeling Analysis

We constructed a dataset of peer-reviewed literature on water security and resilience by exporting the raw results of the final WOS and Scopus searches. Using the Python libraries pandas and numpy, we standardized each file's field names to a standard schema that covers document metadata, authors, keyword fields, and abstracts, tagged the provenance of each record to the database from which it was extracted (WOS or Scopus), and concatenated the exports. After concatenation, we retained only the standardized columns to ensure consistency across data sources.
Records lacking a DOI or that were previously redacted were excluded, and duplicates were removed using DOI as the unique identifier. Document types were normalized to title case and filtered to peer-reviewed and scholarly outputs (articles, reviews, book chapters, data papers, and conference papers). Because we were computationally analyzing semantic content, we only included publications in English. The initial data processing resulted in a final dataset comprising n = 1,771 peer-reviewed publications (DOI). We computed descriptive summaries (document type counts, year range, and availability of abstracts and keyword fields) to profile coverage and completeness.