Data for Research: Text Mining at JSTOR

• How to access JSTOR's content data via a free program called Data for research.
• What you can get from the self-service portion of DfR.
• How to get large (over 1,000 articles) data sets.

Who cares:

• Faculty members who are computer scientists and/or digital humanists. 
• Researchers who want to play with humanities data. 

Data for Research is a free service from JSTOR available to computer scientists, digital humanists,independent scholars and anyone interested in mining data for the purpose of uncovering new trends and patterns. DFR allows users to select and interact with content data in the JSTOR archive such as data from scholarly journal literature (more than 7 million articles) and a set of primary resources (26,000 19th Century British Pamphlets).

Specific web-based tools in the DFR interface:

• a powerful faceted search interface that can be leveraged to define content of interest through an iterative process of searching and results filtering
• word frequencies, citations, key terms, and n-grams utilized for conducting analysis of document-level data
• topic modeling (classification of subject headings at the article level), a powerful tool for content selection and filtering
• downloadable datasets containing word frequencies, citations, key terms, or n-grams associated with the content selected
• visualization tools

The self-service data set options available to the user are limited to up to 1,000 articles.

How to get larger datasets:

In order to get a dataset larger than 1,000 articles, you'll need to contact us and apply.