Combining discourse analysis and topic modeling

What is it about?

This article combines topic modeling, an advanced method for automated text analysis developed in computer science, with critical discourse analysis. By providing an overview or ‘content map’ of the data, topic modeling provides an enriching complement to discourse analysis, aiding discovery and adding analytical rigor. Using this methodological combination, we investigate how Islamophobic and anti-feminist discourses are connected to each other on a large Internet forum, consisting of over 50 million posts. In addition, we apply social network analysis to investigate how these discourses are connected through engaged users.

Featured Image

Why is it important?

This article illustrates how techniques for text mining can be combined with qualitative approaches to text analysis such as discourse analysis. This enables to approach vast amounts of unstructured text, which arguably is becoming increasingly relevant in an age of digital media and Big data.