Skip to content

Schedule + Topics

Sep 20 > Text Mining as a Research Tool in the Humanities & Social Sciences
Ryan Shaw
http://library.duke.edu/events/digital-scholarship/event.do?id=6321
Time: 2:30-4:00 PM
Presentation Slideshttp://aeshin.org/textmining/

Several trusted digital libraries, including Hathi Trust, a large-scale collaborative repository of digital content, have opened their corpus of works to text-mining research. Popular media have taken notice, as seen in the New York Times announcement that “The next big idea in language, history and the arts?” is “Data.”

Ryan Shaw will provide an overview and a critique of text-mining projects, and discuss project design, methodology, scope, integrity of data and analysis as well as preservation. This presentation will help scholars understand the research potential of text mining, and offer a summary of issues and concerns about technology and methods.

Follow tweets for this event at #textdata, #duketext

Sep 27 > TEI: What It Is, and Why You Should Care
Will Shaw
http://library.duke.edu/events/digital-scholarship/event.do?id=6331
Presentation Slideshttp://wsshaw.org/shaw.tei.2012-09-27.pdf

Using the TEI (Text Encoding Initiative) markup language to describe a text can be time-consuming and difficult, but it can also yield remarkable insights into literary and historical works.  It can also facilitate the granular search capabilities and rich editorial apparatus that are important parts of electronic scholarly editions.  This presentation will help explain the development and rationale of TEI and, using examples ranging from the William Blake Archive to new digital humanities projects at Duke, illustrate how scholars have used TEI to analyze single documents to vast electronic archives.

Oct 4 > NVivo as a Tool for Organizing & Analyzing Texts
Charlotte Clark
http://library.duke.edu/events/digital-scholarship/event.do?id=6341
Presentation Storifyhttp://bit.ly/VB3HtR

This presentation will demonstrate how NVivo 10 — a qualitative software program offered free of charge to all Duke users — can be used to organize and analyze large amounts of non-numeric data, including text, audio, video, and image. The presentation will be tailored in part to the interests of the audience, so participants are encouraged come prepared to give a brief (60 second) description of the reasons for their interest in NVivo. In addition to teaching graduate-level qualitative research methods as a Nicholas School of the Environment faculty member, Charlotte Clark is also an Authorized Trainer nationwide for QSR International, the maker of NVivo.

Oct 18 > Whose Words Are These? Legal Considerations of Text Mining
Kevin Smith
http://library.duke.edu/events/digital-scholarship/event.do?id=6351

From a legal perspective, text mining is both simple and very complex.  Simple, because the process itself does not seem to implicate any protected right held by a copyright owner.  But complex because so many ancillary activities, including the publication of an article based on text mining research, involve copyright, technological protection measures and licensing issues that create a thicket of worries.  This presentation will try to clear some of that thicket and organize the remainder into clear steps to take to decide if a text mining project poses any significant legal risks.

Oct 25 > High-Level Text Analysis and Techniques
Angela Zoss
http://library.duke.edu/events/digital-scholarship/event.do?id=6361
Presentation Slideshttp://slidesha.re/YtaUMR
Presentation Storify: http://bit.ly/PEyYfY

Automated text analysis can help researchers explore unusual terms in a large body of texts, the context or the structure of language use, and differences in language use across corpora.  Besides exploring the language or topical coverage of documents, automated approaches to text analysis can also be used to describe and organize the text sources themselves. Categorization and classification of texts benefit from partially or entirely automated text analysis, and features of textual documents beyond bags of words (e.g., entities, metadata, citations) can be used to create meaningful links between the documents themselves. This talk will address such high-level studies of text, where the words themselves fade into the background and serve as an input into systems that allow us to use documents for tasks from retrieval to domain mapping.

Nov 1 > Mapping the Disciplinary Structure of Cognitive Neuroscience through Semantic & Network Analysis
Greg Applebaum & Elizabeth Beam
http://library.duke.edu/events/digital-scholarship/event.do?id=6371
Presentation Slideshttp://bitly.com/PprbRy
Presentation Storifyhttp://bit.ly/RxFF03

Cognitive neuroscience uses functional neuroimaging to connect neurobiological structures and psychological processes. What connections have been made so far, and how can the discipline advance its integrative goal? In this talk we describe a recently developed semantic network approach that maps word co-occurrences in a large corpus of neuroscience studies (i.e., text from abstracts of published papers) to create quantitative maps of the semantic relationships across the discipline. In this manner, we are able to integrate information across thousands of scientific articles and provide a bird’s-eye-view of the discipline that is not otherwise available from individual studies, typical review syntheses, or even coordinate-based meta-analyses. We will describe both the state of the discipline as revealed through the uncovered semantic maps, and discuss how these analysis techniques allow the quantitative identification of concepts that are understudied relative to their importance in the discipline.

Nov 8 > Visualizing Text: Tools and Techniques
Eric Monson
*Location: BOSTOCK LIBRARY 023
http://library.duke.edu/events/digital-scholarship/event.do?id=6381&occur=13841
Presentation Slides: http://slidesha.re/X3mTnm

One of the challenges of introducing technology and visualization into a research project is figuring out how to fit the techniques to the scholarship, rather than the other way around. Questions should drive the tool development, and the technology can inspire and allow new questions to be asked that may have never been considered before. I will present some details of the techniques and technologies I have recently been using to develop two prototype web-based text visualization tools driven by Katherine de Vos Devine’s research into the rhetoric of fashion and intellectual property.

CANCELLED Nov 15 > Mining Oral Histories for Multiple Audiences
Ryan Shaw
Rescheduled for February 13, 2013

Millions of dollars have been spent on digitizing oral histories, preserving records of otherwise undocumented lives and experiences. But a list of digital audio files and transcripts can seem impenetrable and intimidating to all but the most intrepid scholars.
If part of the mission of these projects is to make histories accessible to the public and produce fresh understandings of our past, digitization is not enough. How might computational tools be applied to organizing and presenting these histories in new ways, making them easier to comprehend and connect?

Follow tweets for this event at #OHmining

Dec 6 > How to Read 16,700 Journal Articles: Using Topic Models for Text Mining
Allen Riddell
http://library.duke.edu/events/digital-scholarship/event.do?id=6401&occur=13861

Academic journals record the development of German Studies in the United States over the 20th century. Reading through and documenting trends in tens of thousands of journal articles presents a challenge. In this presentation, I’ll consider alternative ways of reading articles published between 1928 and 2006 in Monatshefte, New German Critique, The German Quarterly, and German Studies Review. One approach, a probabilistic topic model, captures major trends, including the relative decline in articles about language pedagogy and the rise of literary history and criticism.

Leave a Reply