StanceXplore: Visualization for the Interactive Exploration of Stance in Social Media

Research output: Contribution to conferencePaper, not in proceeding

Abstract

The use of interactive visualization techniques in Digital Humanities research can be a useful addition when traditional automated machine learning techniques face difficulties, as is often the case with the exploration of large volumes of dynamic—and in many cases, noisy and conflicting—textual data from social media. Recently, the field of stance analysis has been moving from a predominantly binary approach—either pro or con—to a multifaceted one, where each unit of text may be classified as one (or more) of multiple possible stance categories. This change adds more layers of complexity to an already hard problem, but also opens up new opportunities for obtaining richer and more relevant results from the analysis of stance-taking in social media. In this paper we propose StanceXplore, a new visualization for the interactive exploration of stance in social media. Our goal is to offer DH researchers the chance to explore stance-classified text corpora from different perspectives at the same time, using coordinated multiple views including user-defined topics, content similarity and dissimilarity, and geographical and temporal distribution. As a case study, we explore the activity of Twitter users in Sweden, analyzing their behavior in terms of topics discussed and the stances taken. Each textual unit (tweet) is labeled with one of eleven stance categories from a cognitive-functional stance framework based on recent work. We illustrate how StanceXplore can be used effectively to investigate multidimensional patterns and trends in stance-taking related to cultural events, their geographical distribution, and the confidence of the stance classifier.

Details

Authors
Organisations
External organisations
  • Linnaeus University
Research areas and keywords

Subject classification (UKÄ) – MANDATORY

  • Media Engineering
  • General Language Studies and Linguistics

Keywords

  • text mining, digital humanities, social media, machine learning
Original languageEnglish
Number of pages5
Publication statusPublished - 2017
Publication categoryResearch
Peer-reviewedYes
Event2nd Workshop on Visualization for the Digital Humanities
- Phoenix, United States
Duration: 2017 Oct 22017 Oct 2
Conference number: 2
http://vis4dh.dbvis.de/

Conference

Conference2nd Workshop on Visualization for the Digital Humanities
Abbreviated titleVIS4DH 17
CountryUnited States
CityPhoenix
Period2017/10/022017/10/02
Internet address

Total downloads

No data available

Related projects

Carita Paradis, Andreas Kerren, Magnus Sahlgren, Kostiantyn Kucher, Maria Skeppstedt & Vasiliki Simaki

Swedish Research Council

2013/01/012018/01/02

Project: Research

View all (1)