Research groups

With the NWS Data project, the University of Antwerp wants to investigate how people today, in times of social media, stay up to date with current events. We are a multidisciplinary team consisting of computer linguists (CliPS), data miners (ADM), political scientists (M2P) and communication scientists (MPC). Each research group works on a number of specific research projects within the NWS data project:

CLiPS

At CLiPS, we study which knowledge we can automatically extract from text in the context of political communication. This allows us to study more actors, more texts and more media than would be possible with classical studies. There are three types of information that we extract from text using AI language analysis algorithms: factual information (what the text is about), subjective information (what opinions and sentiments do we find) and meta-information (what can we know about the author of the text, e.g. personality, level of education, etc.). For more information you can mail to: walter.daelemans@uantwerpen.be.

ADM

The Applied Data Mining group focuses on the use of data mining techniques for a better decision-making process. Within this project we analyze social media data (such as Twitter and Facebook) using data mining, text analysis and network analysis to analyze political theories on a fine-grained level. On Twitter we have studied both the interactions between politicians and journalists and the topics politicians tweet about. On Facebook we applied predictive modeling techniques to gain insights in political preference based on Facebook likes. For more information you can mail to: david.martens@uantwerpen.be.

M2P

Old news, new communication: This project looks at the Facebook behavior of politicians. More specifically, the type of news articles that politicians share with their followers on social media and how they do this. The facebook posts of all Dutch-speaking politicians in Belgium were investigated for the period of 1 month. The posts with a news article were selected and coded from this. These posts were compared with the full news agenda to see if politicians use a concept called "selective sharing", or in other words: do politicians mainly share articles that turn out to be positive for them? And how can we explain this selectivity? For more information you can mail to: peter.vanaelst@uantwerpen.be

MPC

MPC approaches the project from a journalistic perspective. We want to use multimethodological research to investigate how the increasing use of web statistics and social media has an influence on the news production process. First, the focus is on the production context. Through in-depth interviews with Flemish news editors, we map out which tools and technologies they use to monitor the behavior and interests of their audience. We also investigate how the insights that editors and journalists acquire through these analytical tools have an influence on the news selection process and the daily functioning of the editors. In addition to qualitative research, we also conduct a survey to gauge the perceptions of individual journalists about the importance and effects of algorithms in the news production process. For example, we want to find out to what extent algorithms have an influence on the gatekeeping process and journalistic news values. Finally, we focus on the news itself: to what extent and on what parameters can we predict whether news stories will score well online and on social media? For more information you can mail to: steve.paulussen@uantwerpen.be