How can we use Natural Language Processing methods to analyse interview transcripts?

Speaker(s):

Shunqi Zhang, Maria Pampaka, Laura Black Zhongyan Chen, Sophina Choudry, University of Southampton

Abstract:

One major challenge of the analysis of interview transcripts is the high amount of time required for manual analysis. In this session, we wish to explore how Natural Language Processing (NLP) methods can speed up transcript analysis with methods such as topic modelling and sentiment analysis. We will share some preliminary results of our application of topic modelling with transcripts also used for earlier published work to start the discussion. Then, we wish to collaborate towards a co-produced journal article that promotes applying NLP methods in social science (and in particular qualitative interviews analyses).. In this session, our aim is to explore NLP techniques for analysing interview transcripts, with two main goals: To gain quantitative insights from qualitative data, and To support and enhance the analysis process itself. We will demonstrate these techniques using material from our previous projects. However, we are also keen to co-produce a journal article with session participants to reflect a shared understanding of how NLP can be applied in broader social science contexts. To that end, we warmly invite participants to contribute in various ways. If they have applied NLP techniques in their own research, we encourage them to bring examples to share during the session. If not, we welcome contributions in the form of thoughts, concerns, or even raw materials that could be analysed during the session. These contributions will help shape the article and ensure it reflects diverse perspectives and practices.