Application of text analytics to analyze emotions in the speeches

Abstract

The paper presents the different aspects of analyzing public speeches of deputies in the Parliament (Sejm) of the Republic of Poland with use of SAS tools for text analytics. A document repository was created based on publicly available transcriptions of speeches for 7th (from Nov 2011 to Nov 2015) and 8th (from Nov 2015 to Jan 2018) term of the Parliament (Sejm). A database contains 440 pdf files with transcriptions of the full-day parliament session. This repository was cleaned and preprocessed, every file was split into a set of personal speeches. As a result, the source data table contains 350 000 records. The aim of the experiment was to check whether automatic analysis of text data is suitable for monitoring the ‘temperature’ of the Parliament (Sejm) debates and the main elements of the rhetoric used.