Developing Analysis Methods for Text Network Analysis of Political Texts

Leader: Kimmo Elo, University of Turku; Partners: AALTO; Collaborator: JYU

Text network analysis (TNA) is a rather new approach in text mining, but has gained popularity when it comes to understanding thematic pathways and text recycling in unstructured document corpora by de- and re-constructing texts. Since TNA can be used for exploring all kinds of unstructured textual data, the foreseen impact is an interesting approach to the exploration, analysis, modeling and visualisation of written and spoken language. Additionally, TNA as a flexible method is well-suited to tackle research questions and problems beyond text mining and topic modeling, as well as to process unstructured texts. UHEL and Aalto have produced the FinnParla knowledge graph, a Linked Data knowledge graph of all ca. 1 million speeches of the Parliament of Finland 1907–2021 interlinked with a related prosopographical knowledge graph of 2800 Members of Parliament and related politicians. We need to discover the key content and trends in text corpora. We will deliver easily applicable tools for TNA-based analysis of unstructured texts, and a sample text network dataset of selected Finnish parliamentary debates.