10

I had the thought that I should run a computation on my document to see things like: How many times I cite each source, how many times I cite each author etc. Then I thought that isn't enough. What if I cite A_foo, B_foo and C_foo, but both A_foo and B_foo cite C_foo. That is worth knowing. A graph of citations between the texts that I am citing will revel interesting information. For example nodes with out any parent, will generally indicate new and original ideas. While they may draw on other things, those things are distant enough from my work that I am not citing them. Nodes that have no links to any others, show that I am bringing in an idea from another subfield, perhaps.

This graph should be computable, I know google scholar maintains a list of almost everything I've cited, and for each thing lists who has cited them.

Has this been experimented with? Are there existing tools for the job? If there isn't I might mess around with some python and put something together.

My references are all in bibtex.

I would like to construct a graph of the citations.

6

Strangely enough, this question was asked on SO a while back with no real good answers. There was a link to this question, which links to PaperScope, which looks promising, although there hasn't been activity on the project since April 2013. I've never used it that tool.

Microsoft's Visual Explorerisn't exactly what you're looking for, but it's related. Hubmed used to have a graph view but I can't find it now, maybe you can.

The problem will be that even going just two nodes out - your citations, and your citations' citations - will include thousands of papers from hundreds of authors. As discussed in the aforelinked SO and SU questions, simply choosing an appropriate way to display the data will be a non-trivial challenge, and making useful inferences from the results will probably be even more difficult. Still, it may be interesting. Good luck!

| improve this answer | |
  • In terms of collecting data from google scholar it does not appear to be very easy - no API last I knew. I've seen other examples where researchers say they download articles from Web of Science and then create the network (see an example graph here). – Andy W Apr 29 '14 at 17:06
  • I don't have access to Web of Science though, so it is unclear what download the articles means exactly (if it involves cleaning out citations from PDFs then that sounds like a difficult task, although from the description it may come as more structured than that). – Andy W Apr 29 '14 at 17:07
  • "The problem will be that even going just two nodes out - your citations, and your citations' citations - will include thousands of papers from hundreds of authors" I only want my citation's citation's If they are also by direct citations. This cuts the nodes right down. – Lyndon White Apr 30 '14 at 0:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.