Co-authorship analysis tool

This software is made to help analyse the connections between publications. Which papers share authors, which authors publish together? What are the timelines? etc.

Robustness goes over precision. The analysis should work with any given bibtex-file that has minimal information. The cleaner the bibtex-file (duplicates, name spellings, ...) the better the output, i.e. more precise. This software is not intended for scientific analysis of text corpora but for understanding connections in selected publication lists.

It's a fork from Simon Carrignon's original lines of code that can be found here. Simon is also on Github.

Prerequisites

The software is developed, used and tested with python3. It requires bibtexparser which can be installed with pip3 install bibtexparser. Other non-standard libraries are networkx, matplotlib and numpy. For best results use a markdown renderer like pandoc to get .pdf-Files.

Usage

./main.py BIBTEXFILENAME

It expects a bibtex-file at FILENAME. A report is written to FILENAME.report.md which can be viewed with a simple text editor. But it looks much nicer when rendered as pdf or html. The example pdf is made with pandoc phd.bib.report.md -o phd.bib.report.pdf.

So far the network has to be visualized with a third party program. Use the following files:

• FILENAME.authorlist.csv
• FILENAME.authorNetwork.csv - network of authors who share a publication
• FILENAME.paperlist.csv
• FILENAME.paperNetwork.csv - network of publications which share an author

To run the example do:

cd example
../main.py phd.bib
pandoc phd.bib.report.md -o phd.bib.report.pdf

Gephi does a pretty good job at drawing networks. The two following graphs were done with Gephi using the output from the example-bibtex.

Warranty

There is none. The software might or might not work for you or even have unexpected effects like data loss or much worse. But I guess it's rather safe to run. In any case: always have a backup at hand!

Authors

• Niko Komin (see Laikaundfreunde)
• Simon Carrignon (see Framagit, on Github or on Twitter).

License

See file LICENSE.