Text analysis of Swedish political discourse
Title: Text analysis of Swedish political discourse
SNIC Project: SNIC 2018/7-49
Project Type: SNIC Small Compute
Principal Investigator: Miriam Hurtado Bodell <miriam.hurtado.bodell@liu.se>
Affiliation: Linköpings universitet
Duration: 2018-09-26 – 2020-10-01
Classification: 50401


Studies of societies, with their cultural and political processes, have been the interest of researchers from different disciplines throughout the ages. However, the access to data on a large scale and the ability to model and extract knowledge from this vast amount of data has not been available to researchers within the social sciences for an equal amount of time. Taking advantage of all forms of data that are able to capture political opinion and changes in the political arena would be essential to represents the complexity of these processes. Using data in the form of text, such as political speeches in the parliament or other assemblies and the politicians social media commentaries to perform text analysis on a large scale is crucial to understand how political opinion evolve in society. Using parliamentary speeches, accessible through Språkbanken, will be used to explore how party frames political issues, how they react to one another and how this might change over time, especially when new specialized parties enters the parliament. Detection of representatives of the different politicians through parliamentary speeches is also a main topic of investigation. We will be using extensions of the traditional topic-modeling approach to capture agenda and framing, and different classification algorithm to assess the predictability of representatives among politicians.