Discursis: a computer-based visual text analytic tool for analysing communication data
Registration closed.
The event is also technically sponsored by the IEEE Hong Kong Section Education Chapter.
Abstract
This workshop presents a tool for discourse analysis, Discursis, which has been developed at The University of Queensland, Australia. Discursis is a computer-based visual text analytic tool for analysing communication data, and is especially useful in discourse analyses of conversational texts. Discursis automatically processes transcribed text to show participants’ individual concept use, and their engagement patterns with other participants around concepts, turn by turn over the entire time-course of a conversation.
In this workshop we will present the basic principles underlying Discursis, along with its strengths and limitations in aiding discourse analysis. Together with workshop participants we will work through a number of examples from popular Australian television interview programs. We will show how conversations can be presented visually, and the kinds of information that can be gleaned quickly to characterise the conversation and focus on key extracts. Finally, we will discuss how Discursis can be integrated with the interpretative skills of the researcher to provide rich interpretations of discourse data.
Course structure
DAY 1
The first day of the workshop will involve participants learning the basics of the software including how to load conversation transcripts, visualise results, and interpret findings.
DAY 2
The second day presents an opportunity for workshop participants to load their own datasets using the software and utilise some of the more advanced features of the software.
Perspective participants
Participants should be reasonably confident with using a computer and know the basics of file organisation, naming conventions (.txt, .csv). If participants can bring pilot data for the second day there will be an opportunity to examine this data. The data format required by the software is comma separated value files, however there will be some support given to demonstrating how to manipulate other formats into this format during the workshop. The basic csv format required is for a single column with the transcribed text organised in turn order row-by-row, and other metadata organised in separate columns (speaker ID, turn length, physical gestures, etc.):
Text,Speaker
“Hi, how are you”,John
“Good John, I hope you are well”,Nancy
About the speaker
Dr Daniel Angus completed a PhD in computer science at Swinburne University of Technology in 2008, shortly before joining The University of Queensland as part of the ARC Thinking Systems special initiative under the mentorship of Dr Andrew Smith and Prof. Janet Wiles. From 2008, Angus, Smith, and Wiles pioneered the development of the Discursis computer-based visual text analytic tool, a methodology for assisting the analysis of communication data. In 2012 Dr Angus began a strategic initiative in communication technologies between the School of Journalism and Communication and School of Information Technology & Electrical Engineering. In 2015 Dr Angus was appointed as lecturer in the newly formed School of Communication and Arts, with research and teaching activities focussing on the development and evaluation of advanced computational methodologies for journalism and communication.
Related event
Summerfest 2015
Stimulating and Incubating Interdisciplinary Research on Learning
