Blog

Visualising tag data (from Zotero?)

I’m currently writing a methodological justification. For that, it crossed my mind that a nice appendix would include articles that have used various methods, as well as some little descriptors (in this case I was primarily thinking epistemological positions). That data – methods, epistemology – could be stored as a set of tags.

I want to try and avoid ‘stacks’ (as per delicious/folders, etc.) because what I’m interested in is representing how methods have been linked together, in such a way as to allow access to the articles (i.e. you can ‘expand out’) while also indicating the ‘structure’ of the field.

First off, I wondered whether I could use some tool to tag things as I save them, and have it auto-link those that share some tag as I went along.

However, given that I already use Zotero, which as a perfectly good tagging system, that seems unnecessary.

So, I’d like it to look like:
http://mashe.hawksey.info/2011/10/guardian-tag-explorer/
Or http://blog.ouseful.info/2011/01/27/tags-associated-with-other-tags-on-delicious-bookmarked-resources/

Access to libraries in Zotero is pretty easy; you can see them online, (e.g. mine is here
https://www.zotero.org/goodwin57/items) and you can export them in various forms including – .rdf, MODS, BibTeX, COinS, RIS, TEI.
And there is an api.

Also, someone else seems to have written something that will capture some of the zotero data from an exported file (less desireable than using the api I reckon).

Problem is, although I’m not adverse to a bit of code tweaking, I am fundamentally a moron. So I don’t really have a clue where to start with this stuff. I do, however, think it would be useful – although I’m willing to be shouted down on that too.

So! Any ideas, solutions, thoughts on the concept in general?


Print pagePDF pageEmail page

This Post Has 3 Comments

  1. Tony Hirst says:

    Assuming that I can just grab an exported version of a collection, I’d start by looking at the API docs – http://www.zotero.org/support/dev/server_api/read_api

    Docs say: “Requests for data in a specific library begin with /users/ or /groups/. User IDs are different from usernames and can be found on the API Keys page and in OAuth responses.”

    I’m not logged in as you so need to find another way of finding the userID. Guessing that Zotero uses its own api, I go into my browser developer tools and see what URLs get loaded when I click through to one of your collection pages. I see a URI of the form https://www.zotero.org/api/users/2449/collections/E6VP3WPG/items/top?start=0&limit=25&order=title&content=json&_=1328289947196 so I guess your userID is 2449

    If I click on your PhD folder the page that’s loaded is https://www.zotero.org/goodwin57/items/collectionKey/B3SFKWM5

    So looking at the API docs, these URIs should give the items and subcollections of that collection:
    https://api.zotero.org/users/2449/collections/B3SFKWM5/items
    https://api.zotero.org/users/2449/collections/B3SFKWM5/collections

    API docs suggest results are paged but there doesn’t seem to be any statement of how many items there are in a collection, which is faff. It means you need to load batches of results in at, say, 50 a time (the default), and then if you get less than fifty results in a page, increment the index of the next result you need and grab another page (so the first page set start=0 (I assume it starts the index count starts at 0), second page set start=50, third page start at 100 and so on. If a results page contains less than 50 results, you know you’ve loaded all the pages. (If you go too far, you get no items in the results page).

    So, what I would do is…erm… not sure…Scraperwiki and Python maybe… been a long day.. decided I don’t like Zotero API and don’t want to give up my Friday night learning about it…;-) Maybe over the weekend…

  2. sjgknight says:

    Wow… things I want to learn about and that make me scream in the afternoon. Ironically, I’m only doing this because I shattered after a door slamming woke me up at 3.30am n unable to sleep again! Perhaps not the best state to be looking in…

    Anyway, at the moment I’m trying to see if I can create a CSV that looks ok to work in Gephi (which I’ve never used before)… I’ll run through some tutorials too…
    Thanks for the tips (direct and indirect) thus far!!

Leave A Reply





%d bloggers like this: