AI Literacy in Medical and Health-Professions Education (2020–2026): Data and Code.

This repository accompanies the bibliometric scoping review *"Artificial Intelligence Literacy in Medical and Health-Professions Education: A Bibliometric Scoping Review of a Rapidly Emerging Literature (2020-2026)."* It contains the analyzed corpora, the theme assignments, and the code that reproduces the figures and tables. Contents data/ ai_literacy_corpus.csv Parent AI-literacy corpus (n = 12,326) ai_literacy_medical_corpus.csv Strong-signal medical subset (n = 1,141) ai_literacy_medical_clustered.csv Clustered set with theme labels (n = 957) ai_literacy_medical_threads_summary.csv Theme-level summary table (8 themes) code/ 01_retrieve.py Retrieve the corpus from OpenAlex (needs OPENALEX_API_KEY) 02_filter.py Apply the strong-signal medical filter 03_cluster.py TF-IDF -> SVD -> k-means; silhouette + stability diagnostics 04_figures.py Regenerate figures and the theme-summary table requirements.txt Pinned Python dependencies docs/ supplementary_materials.md Query strings, filter rules, parameters, diagnostics column_dictionary.md Definition of every data-file column Reproducing the analysis pip install -r code/requirements.txt # (optional) re-retrieve from OpenAlex - reproduces the parent corpus: OPENALEX_API_KEY=your_key python code/01_retrieve.py --out data/ai_literacy_corpus.csv # apply the medical filter: python code/02_filter.py --in data/ai_literacy_corpus.csv --out /tmp/med.csv # clustering diagnostics (silhouette scan, seed stability): python code/03_cluster.py --in data/ai_literacy_medical_corpus.csv --out /tmp/clust.csv # regenerate figures and the theme table from the frozen clustered set: python code/04_figures.py \ --clustered data/ai_literacy_medical_clustered.csv \ --full data/ai_literacy_medical_corpus.csv --outdir figures Reproducibility notes (please read) The CSV files are the authoritative record. They freeze the exact corpus, subset, and theme assignments reported in the paper. 01_retrieve.py queries a live database; counts may drift as OpenAlex is updated. The provided ai_literacy_corpus.csv is the version used. 02_filter.py recovers >99% of the published medical records (1,137/1,141) but is slightly over-inclusive (it flags ~1,415 records with the documented term list). The frozen ai_literacy_medical_corpus.csv (with its strong_signal column) is the authoritative published subset; use it for exact counts. 03_cluster.py reproduces the subset size (957), the SVD variance (0.287), the silhouette scan, and the seed-stability ARI (~0.43) exactly. Because k-means on sparse text is only moderately stable, exact per-paper cluster assignments are not reproducible run-to-run; the frozen ai_literacy_medical_clustered.csv carries the final cluster indices and reviewer-assigned theme labels, and 04_figures.py uses it to regenerate the published figures and tables exactly. Data provenance and license Bibliographic metadata are derived from OpenAlex (https://openalex.org), which releases its data under CC0. Abstracts were reconstructed from OpenAlex inverted-index fields. This deposit is released under CC BY 4.0; the code is released under the MIT License (see LICENSE). Citation If you use these materials, please cite the accompanying article and this deposit (see CITATION.cff).