Interactional data models for topic and sentiment analytics (2026-2029)

Text analytics can be used to assess and evaluate what people are talking about (topic) and how they feel about it (sentiment, stance). However, current computational approaches for analysing topic and sentiment are limited by under-specification and ad hoc incorporation of interactional context. The aim of this project is to develop formal interactional data models that address the dialogic properties of communication, which can then be used as a blueprint for re-engineering topic and sentiment analytics. Grounded in an approach to data modelling as research practice, this project expects to develop novel data models for analysing topic and sentiment in common modes of communication across different interactional contexts. These formal interactional models can then provide the basis for engineering next-generation interactional topic and sentiment analytics that fully leverage the affordances of data-driven, real-time analytics for addressing human behaviour and influence challenges.