1

I would like to know how many different authors have posted something on the different categories of the arXiv. Are there any statistics somewhere about this?

To be precise what I'm looking for:

If there are no such statistics, a pointer to a search engine that provides such data for different scientific fields would be a great help.

Improve this question
5
  • 2
    The arXiv supports the Open Archive Initiative arxiv.org/help/oa/index so in principle this is just a programming question. (Request all articles in Astrophysics, extract the author names, sort, kill duplicates, count.)
    – Willie Wong
    Apr 10, 2015 at 13:44
  • Also, it appears that the arXiv analytics website (note: not officially affiliated with arXiv) may perform visualizations of author contributions. But I haven't been able to get it to work in my browser yet, and as far as I can tell it may not have been implemented.
    – Willie Wong
    Apr 10, 2015 at 14:28
  • @WillieWong Can you turn your first comment into an answer, so that I can vote it up?
    – jakebeal
    May 22, 2015 at 5:19
  • @jakebeal: done.
    – Willie Wong
    May 26, 2015 at 7:22
  • @WillieWong ... and done!
    – jakebeal
    May 26, 2015 at 12:45

1 Answer 1

Reset to default
4

(Now an answer by request.)

The arXiv supports the Open Archive Initiative so in principle this is just a programming question for

  • Requesting and downloading the data for all articles in Astrophysics and dumping it locally in a usable format.
  • Extracting the author information.
  • Culling the duplicates.
  • Counting the resulting entries.

If the first step is done right (by putting the information in a suitable database format) the other three will be more-or-less trivial.

Note however that name duplication would be a problem that is hard to solve, especially when some articles are listed with a "last name, first initials" format. Similarly authors who use more than one version of their names on arXiv (I'm guilty of that myself) can cause some inaccuracies to the computation outlined above. The mathematicians solve this "manually" by employing 13 people and making this part of their jobs, and I don't know if there can be a good programmatic solution until at least wide adoption of some sort of author ID (like ORCid) system.

Improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .