The new version of revtools is up on CRAN and GitHub now. This is a big change from the previous version, and so while there are many useful new features, there are also some changes to the API. Consequently, you should be careful about upgrading if you are using revtools for an ongoing project. Most importantly, the save function in start_review_window() is not reverse-compatible, so updating will mean you lose the ability to reload your progress from earlier versions.

If you want a complete overview of how revtools works, you should check out the links in the header; but if you just want to see what’s changed since the last version, here is a quick summary:


  • new colour scheme and logo for shiny apps
  • new apps for screening duplicates, titles or abstracts
  • import functions always return a data.frame by default
  • more versatile duplicate detection, including new string matching algorithms
  • improved topic model control to allow users to interactively select included text

New functions

  • apps: screen_titles, screen_abstracts & screen_duplicates
  • fuzzdist: Fuzzy string matching based on the Python library 'Fuzzywuzzy'
  • revwords: Stopwords, but including words that describe numbers

API changes

Note old function names have been retained where possible, but may be deleted from later versions.

  • all function names are now lower case: make_DTM is now make_dtm; run_LDA is now run_topic_model.
  • start_review_window is now screen_topics for consistency with other apps
  • read_bibliography now returns a data.frame by default, but returns an object of class 'bibliography' if return_df = FALSE. This function also accepts .csv files now.
  • arguments to find_duplicates are different
  • extract_unique_references returns a vector, not a data.frame