Journal article
CarSenToGram: geovisual text analytics for exploring spatiotemporal variation in public discourse on Twitter
Cartography and Geographic Information Science: Special issue on frontiers of geospatial data science from the joint UCGIS Symposium / AutoCarto 2018 Conference, Vol.46(1), pp.57-71
01/02/2019
DOI: 10.1080/15230406.2018.1510343
Abstract
Assessing the impact of events on the evolution of online public discourse is challenging due to the lack of data prior to the event and appropriate methodologies for capturing the progression of tenor of public discourse, both in terms of their tone and topic. In this article, we introduce a geovisual analytics framework, CarSenToGram, which integrates topic modeling and sentiment analysis with cartograms to identify the changing dynamics of public discourse on a particular topic across space and time. The main novelty of CarSenToGram is coupling comprehensible spatiotemporal overviews of the overall distribution, topical and sentiment patterns with increasing levels of information supported by zoom and filter, and details-on-demand interactions. To demonstrate the utility of CarSenToGram, in this article, we analyze tweets related to immigration the month before and after the 27 January 2017 travel ban in order to reveal insights into one of the defining moments of President Trump's first year in office. Not only do we find that the travel ban influenced online public discourse and sentiment on immigration, but it also highlighted important partisan divisions within the US.
Details
- Title: Subtitle
- CarSenToGram: geovisual text analytics for exploring spatiotemporal variation in public discourse on Twitter
- Creators
- Caglar Koylu - Geographical and Sustainability Sciences, University of IowaRyan Larson - Computer Science, University of IowaBryce J Dietrich - Political Science, University of IowaKang-Pyo Lee - Iowa Informatics Initiative, University of Iowa
- Resource Type
- Journal article
- Publication Details
- Cartography and Geographic Information Science: Special issue on frontiers of geospatial data science from the joint UCGIS Symposium / AutoCarto 2018 Conference, Vol.46(1), pp.57-71
- Publisher
- Taylor & Francis
- DOI
- 10.1080/15230406.2018.1510343
- ISSN
- 1523-0406
- eISSN
- 1545-0465
- Language
- English
- Date published
- 01/02/2019
- Academic Unit
- Research Services; Political Science; Biostatistics; Business Analytics; Geographical and Sustainability Sciences; Public Policy Center (Archive); Center for Social Science Innovation
- Record Identifier
- 9983982715502771
Metrics
33 Record Views