next »

[tee_bee]

All,

I apologize if this has been covered before. Here's my situation. I am embarking in some research in a part of my field with which I am not familiar, but which crosses over to the literature with which I am very familiar.

So here I sit with a bunch of JSTOR references, and I am reading things like review articles and studies. They develop ideas from other authors, and then advance their arguments. What I need to do is figure out how to take notes on

1. What the article in front of me says
2. Who they cite to support the idea

Then, what I would like to do when all is said and done is

3. Be able to forward map and backwards map the citations to see who is citing whom, how the ideas are converging or diverging, etc.

Example:

In front of me I am reading a 1991 article by Smith. For one particular idea, Smith cites Jones, Able, Baker, Charley, etc., as supporting the claim. Then, in later papers, others cite Baker, who turns out to be a big wheel in the field (I could already tell this from my first perusals of the articles). I am trying to build, in a sense, a mental map of the fields I am reading about. I tend to think very spatially/temporally, and when I can "visualize" ideas and their authors, and their connections, in time and space (where "space" could be a continuum of ideas, or intellectual conflicts, or whatever) I can really understand what I am reading and how it fits together. I am thinking something like network analysis might be similar--and what I am doing is sort of a bibliometrics problem, but I am trying to track ideas, not just authors.

[For those who are familiar with old law books, what I am trying to do, I think, is "Shepherdize" the literature I am reading, but both forwards and backwards.]

The problem--I tried to set up something like this in an MS Access database, but it seems illsuited to this purpose.

Any ideas out there? This is something that has vexed me for over 20 years now. Any systems, software, shareware, web-based stuff? Let me know if this is utterly unclear, and I will try to clarify what I am looking for. Thanks!

[locutus]

I think there are databases that will do something like this for you. WebofScience, WebofKnowledge maybe? I've certainly seen online citation maps.

Is what you want to do somewhat similar to this: http://chronicle.com/forums/index.php/topic,64318.0.html

[tee_bee]

This link is good, thanks! But web of science won't work for me--its coverage of social science ideas is poor, and its main unit of analysis is the citation, not the ideas themselves. At least, that's my impression. Am I using it wrong?

[locutus]

This link is good, thanks! But web of science won't work for me--its coverage of social science ideas is poor, and its main unit of analysis is the citation, not the ideas themselves. At least, that's my impression. Am I using it wrong?

I've only seen it for citations. I thought maybe it would have something for keywords too.

That other thread is about my project that I think is similar to what you are trying to do. It sounds like we have a similar motivation of trying to get a "mental map" of the literature. I don't really have any special software that does everything. I'm using Mendeley to catalog papers. That works well and is free. It doesn't create any visual representations though. I have futzed around with MATLAB and that has produced a few things. I think that my situation is a bit different than yours though. I am working with a set of papers that are pretty similar, so I can set the characteristics to numerical values,  each paper ends up as a vector and then run a cluster analysis on all those numbers.  So the "ideas" are created using a bottom up approach.

Your idea sounds interesting...I have to think about it some more.

[bread_pirate_naan]

Look at Zotero.  You can download citations and take any number of notes, attache images and other text files and select or create any set of searchable terms for each bib record and each child not within that record.

You can then search any of those terms, from a bibliography sorted by author/date/title (whatever) and generate a text report that does pretty much exactly what you are trying to do. (chronological reports on your selected terms will probably work best)

You can't generate the visualizations you want, but...it's good.

If you have govt or industry ties, find out who is using Documentum.

[tee_bee]

Thanks, Locutus. Your ideas are great, and the other forum thread has been very helpful.

BPN: I've tried Zotero, and maybe it's just me, but I cannot seem to make it do anything more for me than Endnote does. And I don't use Firefox anymore (plug-ins make FF dog slow), and I want to stick with Chrome. But if Zotero works better in the end, this is a reason to go back to FF.

Documentum sounds like a good idea. And I am going to try Mendelay again. Since this is a shared project with several colleagues, the library sharing features seem good. And apparently Zotero works and plays well with Mendelay

[bookishone]

If you want to make the relations visual, then maybe a concept-map type application would work. Something like OmniGraffle or NovaMind.

http://www.omnigroup.com/applications/omnigraffle/

http://www.novamind.com/

I would think that all the colors/shapes in NovaMind would distract from the underlying patterns, myself, but I can also see that it is nice to have that additional freedom to customize the connections.

The other possibility would be Tinderbox, which is more than just a mapping tool. This is a very powerful note-taking program that has outline view, "tree" view, and map view as well as others -- very robust, but also a steep learning curve. People who learn it are converts for life. I haven't had time to learn my way around it yet:

http://www.eastgate.com/Tinderbox/

The advantage of this type of solution is that I think you should be able not only to indicate relation but also the type and direction of that relation (strong cite, passing cite, mutual cite, etc.). I should add that I don't currently use any of these programs, but I fantasize about getting organized enough to do so... maybe after tenure.

Seems to me that the kind of map you have in mind would be very useful, but you want to be careful not to let yourself get bogged down into creating the perfect and most comprehensive map of the literature, which would be tempting. Find a solution that integrates well with basic note-taking if possible.

Maybe others on the fora have used these programs or others and can weigh in.

[tee_bee]

I'd love to use Omnigraffe--unfortunately, I run Windows XP. Maybe I will tinker with Visio again to see if I can make this work.

[bread_pirate_naan]

BPN: I've tried Zotero, and maybe it's just me, but I cannot seem to make it do anything more for me than Endnote does. And I don't use Firefox anymore (plug-ins make FF dog slow), and I want to stick with Chrome. But if Zotero works better in the end, this is a reason to go back to FF.

I have an irrational dislike for endnote from a past life.  And I don't like chrome at all.

Documentum sounds like a good idea. And I am going to try Mendelay again. Since this is a shared project with several colleagues, the library sharing features seem good. And apparently Zotero works and plays well with Mendelay

Mendelay would be great if it networked documents to a single server the way zotero does.  I float around too much and don't want all the docs tied to a single computer.  Any online bibliography does roughly the same thing with decent proxy service.  I do not see the ability to import text from wikipedia as a selling point.  It also has fewer citation styles than zotero.  Most of the other point and click things it does came hardwired in my brain.

I do like that it is relational, and will try it out.  While I am in a book field, I can see how it would be fantastic for scientists of any stripe. The text marking too is fantastic.  If the next release actually makes that collaborative, it will be indisputably superior, but why would they do that and not allow the documents themselves to follow the user?  Counter-intuitive.

[bread_pirate_naan]

I tried it. 

The import system is weak.  It collapses all notes attached to a record into one file. There is only a single note field attached to any bib record.  All the tags are imported from the bib record and all the note tags of individual notes disappear.  If you are organized in note taking this is bad. You can only make notes in a side panel, so if you like to write in the big window in zotero, your are SOL. 

One weakness, it doesn't indicate what sites it operates on.  Zotero does this with a nice icon in the address bar that recognizes what sort of media/bib file it is.  You have to remember where Mendeley functions and then use an import tab in the toolbar.  It does not function with my uni library interface, weakening Mendeley's use value to somewhere near zero.

Far inferior to zotero for the seriously organized or nuanced library/index type thinker/writer.  The good part is the importation of whole documents.  I might use it for that, as one can run both programs easily, but it's not a good use of time.  I'd rather support and help develop Zotero, which will always be free.  Other than that it is mainly a weak zotero with a social networking function attached.  This is as much a plus for me as wikipedia import function and no relationship to my library.  Expect there to be fee for servi9ce functions soon.

Thumbs down on Mendeley for qualitative, detailed, budget conscious scholars, or current Zotero users.  Way down.

[locutus]

Interesting. I actually tried Zotero out before I got Mendeley. I just couldn't quite get how it was meant to work and gave up. Mendeley seemed a lot easier, though it does have issues (in their defense it has improved quickly and will hopefully continue to do so). I'm not a serious not taker, I just want a nice index for all my papers with the ability to tag and sort them.

[bread_pirate_naan]

Interesting. I actually tried Zotero out before I got Mendeley. I just couldn't quite get how it was meant to work and gave up. Mendeley seemed a lot easier, though it does have issues (in their defense it has improved quickly and will hopefully continue to do so). I'm not a serious not taker, I just want a nice index for all my papers with the ability to tag and sort them.

The best thing about zotero is that I can note any little thing that relates to any of the number of topics I think about really hard and ideally, in depth (at last count, my advisor is running somewhere between 4-6 books).  At any given time I can aggregate all the notes with a single tag, from any topic and all I have to do is click a nice little report tag and it generates a firefox tab with a full report -- including all the tags attached to that note. (for better cross referencing)  As Zotero is super intuitive, it keeps me from generating a folksonomy of random tags and predicts/shows all taxonomic terms with similar spelling -- the second I start to type.  And you can give each note multiple tags.

So, say I am working on Potatoheads.  When I search "potato" in the entire library, it not only aggregates every title in my library with that Library of Congress keyword, it also highlights any note in any bib record with that tag, too.  Now I can start a potato file, which is merely another sort of tagging system, not a duplication of any sort.  Then I can search that library or the whole library or any other topic folder for "eyes" or "hats" and start to gather everything I have ever thought about the topic for my much needed paper on the deeper political significance and economic impact of potatoheads wearing top hats.

I can also select only a certain portion of the records from that list to generate my report.  It is super nuanced.

If you are crazy, this makes working very easy, because you have so many deep thoughts about potatoheads that reading one's notes for one hat style and leaving out the deeper power of a variety of shoe styles for another go around is dumb.  If you've been reading and making good notes, you can remember to tag what page the shoe references appear on, as well as your deep random thought about shoes, while you are reading and working on hats.  Then you have another paper, or even another chapter of your forthcoming book (proposal).  Potatoheads in space and undersea potatoheads are another matter all together, but if you have used zotero well, your papers may start to write themselves, more or less.  

You have to be invested in the tool and be willing to type your notes.  It's like high tech marginalia-plus.  If they had a highlight and copy tool like Mendeley claims, but I wasn't able to operate, Zotero would be the greatest thing ever. Except for those poor longhand writers.  They're out in the cold.

I can also import my seminar notes from Google docs as a complete pdf(I take all my lecture notes in google docs now), or chew them up and tag them in smaller pieces.

[collegekidsmom]

Web of Science does have an "analyze" feature that will create a citation map with backward and forward mapping for an article. First, the coverage of your field by Thomson Reuters sounds poor, and the visual mapping is for a single citation. For fields covered well by WOS, you might be able to work with it in the way you describe and use EndNote web alongside. Web of Science is also a good subject search tool, but only for the fields it covers well, of course. It is not just for citation searching.
The whole idea reminds me of Otlet's Mundaneum idea, except on a smaller scale.

[locutus]

http://www.histcite.com/

Found this buried in my bookmarks.

[tee_bee]

I'm back, treading water. Thanks for all the advice. Web of Science is sort of cool--but not very helpful for social scientists, unless something has changed in the last six months. I will try it again. It has improved in the social sciences as Thompson has tried to improve its handling of impact scores.