It's looking like my #workflow for a #citation project is:

-Copy text citations from diss/theses into snowelm.com/~t/doc/tips/makebi
-Import resulting Bibtex to #Zotero
-Export .ris from Z to #Mendeley (which won't take the bibtex above)
-Run M's review/search procedure to verify/correct citations and/or manually fix as necessary

then either
-Export back from M to Z
-Export from Z to .csv

or
-Export from Mendeley to R :rstats: (RefManageR looks promising)

This seems...cumbersome.

text2bib.economics.utoronto.ca is another possibility for the initial conversion, but it requires review of each citation during processing, which isn't practical for this project (which is to get citations from multiple diss/theses into an analyzable format, so we're talking 100s of citations)

Another useful site, but not for this project: excite.west.uni-koblenz.de/exc

References came out messy, but potentially usable. It would probably work better from publisher-formatted PDFs than scanned dissertations

Yet another similar tool to add to the collection.
scholarcy.com/bookmarklets/
References in bibtex or ris from PDFs on the web

Annnnd....just to have all of these in one place:

For finding DOIs: search.crossref.org/references

And it does work with lists, as long as you format to one ref per line

@delibrarian We've found that sometimes the first match with that search is not the one we want to match on. So I'm not using that particular tool. CrossRef search preferences recall.

@ronallo I've found there is always error. For instance, many of these tools preference or only process journal articles, so book citations get matched to book reviews. Can't get rid of human intervention yet. I haven't used this one enough to get an idea of % accuracy.

@delibrarian Yes, I have student workers doing this work. We've found the same issues with focus on journal articles. We'll be dealing with a lot of citations to technical reports and other unpublished works. Just trying to find something that will help them power through many more enhancements/corrections.

@delibrarian So far underwhelmed by the options and how they have to be strung together for a full solution.

Still trying to decide whether cobbling different tools together would cost less than us trying to develop the tool I really want, especially since a lot of what we're doing is a one off process, just that we're doing a lot of it right now and want to get it done quickly.

@ronallo One of the reasons I posted to begin with was in the hopes that someone else had already figured this out. I don't mind linking tools, but having to do extra processing between them is a pain. Grumble: Isn't this what we have standards for?

Unfortunately, it seems we're all in the same place right now. Tools are in the works, or are pilot projects. A lot of tedious work that ought to be automate-able that isn't yet.

Follow

@delibrarian Right. That's why I posted too. Surprised that some of this wasn't a more solved challenge. At least now there are some utility services like CrossRef available for some things.

Part of the problem I have is that BibTeX isn't really a standard. I'm going to be looking some at how different systems and libraries specify BibTeX categories/types.

Sign in to participate in the conversation
code4lib.social

code4lib.social is one server in the network