ReviewPaper_DataQualityForMLinMedicine

Supporting code and information for review paper: The METRIC-framework for assessing data quality for trustworthy AI in medicine: a systematic review

The following files are provided:

• Creating literature database
  • "get_unique_xlsx.py"
    • python script for concatenating and deduplicating raw database results
• Plotting Figure 2
  • "fig_2_dotplots.R"
    • R script for generating Figure 2 of the manuscript
  • "provenance_corpus_papers.xlsx"
    • data required for Figure 2

Script for concatenating database results and deduplicating

Run get_xlsx.py

python3 get_uniqe_xlsx.py --path_acm_enw acm.enw --path_pubmed pb.xlsx --path_webofscience wos.xls --path_concatenated cc.xlsx --path_concatenated_unique_doi uniquedoi.xlsx --path_concatenated_unique_titles uniquetitles.xlsx --path_concatenated_unique_doi_unique_titles uniquedoiuniquetitles.xlsx

Figure 2 - Corpus records sorted by publication date + grouped by categories and (non-) life science

1. Open fig_2_dotplots.R (e.g., in RStudio)
2. Edit the path to your directory conatinint the script and provenance_corpus_papers.xlsx.
3. Execute the code to reproduce Figure 2