Paper Machines
Posted on January 22, 2014 in Uncategorized by Eric Lease Morgan
Today I learned about Paper Machines, a very useful plug-in for Zotero allowing the reader to do visualizations against their collection of citations.
Today Jeffrey Bain-Conkin pointed me towards a website called Lincoln Logarithms where sermons about Abraham Lincoln and slavery were analyzed. To do some of the analysis a Zotero plug-in called Paper Machines was used, and it works pretty well:
- make sure Zotero’s full text indexing feature is turned on
- download and install Paper Machines
- select one of your Zotero collections to be analyzed
- wait
- select any one of a number of visualizations to create
I am in the process of writing a book on linked data for archivists. I have am using Zotero to keep track the books citations, etc. I used Paper Machines to generate the following images:
a word cloud where the words are weighted by a TF-IDF score
a “phrase net” where the words are joined by the word “is”
a topic modeling map — illustration of automatically classified documents
From these visualizations I learned:
- not much from the word cloud except what I already knew
- the word “data” is connected to many different actions
- I have few, if any, citations in my collection from the mid-2000’s
I have often thought collections of metadata (citations, MARC records, the output from JSTOR’s Data For Research service) could easily be evaluated and visualized. Paper Machines does just that. I wish I had done it. (To some extent, I have, but the work is fledgling and called JSTOR Tool.)
In any event, if you use Zotero, then I suggest you give Paper Machines a try.
