Utilising Textual Data in Crime Analysis: Insights from the Sex Market and Public Reactions to Crime on Twitter

Text is often used for qualitative research, but it hasn’t been used much for quantitative research. This talk will show how we can use automated text analysis in crime research. We will start by explaining why text is becoming a useful source of information in this area. Next, we'll talk about different natural language processing techniques for understanding large amounts of text data. We'll pay special attention to a method called topic modelling and show two examples of how it's used. The first example is about analysing 72,966 anonymous online reviews from clients of sex workers, written between 2002 and 2020. The second example presents an analysis of crime-related news articles (N = 56,023) and their associated replies and quotes on Twitter (N = 334,827) in the U.K. We'll wrap up the talk by discussing the strengths and limitations about using text for crime research, along with strategies to address the shortcomings of this approach.

Presented by Luca Giommoni, Reader in Criminology at Cardiff University, and Pablo Ezquerra Silva, PhD candidate and Lecturer in Quantitative Social Sciences at Cardiff University.

This seminar will take place in-person. If you're unable to attend in-person please contact qstep@exeter.ac.uk

Book your place here.