A Python script that uses the crossref API to guess the gender of every author in a bibtex file. Inspired by this tweet.
Assuming gender from someone's name can be very inaccurate, and this methodology is unable to account for non-binary people. These results should be used as a starting point for discussions around gender diversity in academic reference lists. For more information on the gender identification methodology, please refer to this page, which outlines the algorithm used for the gender classification.
- Python 3
Clone or download this repo and install the dependencies using:
pip install -r requirements.txt
If you are going to be using this code intensively, please replace my email address in the file called EMAIL with your own. This will allow Crossref to better track who is using their service.
The tool is run from the command line using a single argument, the filename of a bibtex file you wish to analyze:
python gender-refs.py references.bib
Running the code on the provided example bibtex file should give results that look like this:
Found 129 references.
References without a DOI cannot be processed.
Unknown: 19
Mostly male: 3
Male: 195
Female: 29
Androgynous: 7
Mostly female: 2
Total authors processed: 255
If you have suggestions for how to improve this tool, open an issue, and I will try and resolve it, or feel free to open a pull request and fix it yourself!