4

Given a (rather large) list of authors, what is the best way to find the number of publications by each author? The big problem here is author ambiguity (which John Smith, etc.) - the author list we have doesn't even include middle initials! What we do have is the full name of a FDA Committee on which they have sat, which has some useful key terms (e.g., arthritis).

Ideally, some sort of API/automated search could be used given the large number of authors! I've heard about some databases (eRA Commons, for example) but these don't seem to have a visible search page.

What is the best way to figure out publication number given these constraints?

  • Just for clarification: do you only want to search in/using PubMed, or are other databases okay? academic.research.microsoft.com seems to do an okay job separating the large number of Chinese name collisions. (Though of course, it is not perfect.) – Willie Wong Apr 10 '13 at 15:40
  • Have you thought about emailing them and asking for a CV? Pubmed provides the ability to search by affiliation which might help, but it is really designed to find articles and not authors. Scopus's affiliation filter might be better since it is designed to find authors (as well as articles). – StrongBad Apr 17 '13 at 14:35
2

This is a little bit shaky, but let's give it a try. Apart from a programmatic solution where you would have to write your own web-scraping scripts which would construct appropriate URLs for each authors and extract their publication statistics, there might be another, though still quite tedious and technically involved approach.

Harzing's Publish or Perish program (PoP) has a feature called "Multi-query center". It allows you to create a set of queries and then generate publication/citation statistics for each of them in bulk. It is meant to periodically re-generate the citation statistics for a set of queries so that you do not have to laboriously write the queries again and again. Now the technical steps towards your solutions would be the following:

  1. produce a set of queries corresponding to the list of authors you are interested in;
  2. update all these queries at once by a single button click;
  3. save the statistics for all the queries e.g., into a CSV file; and
  4. extract the numbers you are interested in in your favorite spreadsheet program.

The most technically involved step is to produce a set of queries corresponding to the list of authors you have. If the list is not too long (e.g., up to 50), the easiest method would be to enter the queries manually. In the case you have a very long list of authors (several hundreds), then I would try to generate the set of queries programatically. The PoP program stores the queries you enter in it in a file Queries3.xml in the corresponding Program Files folder. The queries are organised into folders, the XML is easy to read. Producing a correct XML file with the queries is a little bit try-error process, but you can easily succeed when you create an empty query with an authors and then using a small script (or you favourite text editor/spreadsheet) copy the query entry and only change the Author attribute of the PoPQuery element.

Good Luck!

1

My answer is a supplement to walkmanyl's answer.

Can I suggest you to go through this paper? It is a very nice paper which gives you an algorithm to disambiguate author names and can be used in conjunction with the method described above.

Anecdotally, I know that it has been used in a number of scientometric/bibliometric studies already.

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.