Exploratory Models in a time of Big Data

Abstract

This paper aims to trigger discourse about the emergence of a new type of social scientific model — Exploratory Models — which draw on Big Data, computer modeling and interdisciplinary research to tackle complex social scientific processes. First, we define Exploratory Models referring to Batty and Morgan and Morrison. We then present changes to the traditional modeling paradigm. We show how Exploratory Models circumvent challenges related to the idiosyncracy, self-reflexivity and acceleration of social phenomena, which limit predictive effectiveness of traditional models. We show that Exploratory Models are better equipped to tackle complex problems due to their capacity to process heterogeneous datasets. Having established that Exploratory Models are predominantly problem- and data-driven, we emphasize that scientific theory is indispensable to their progress. Finally, the development of an integrative platform is suggested as a way of maximizing the benefits of this approach. Discussion concludes by flagging areas for further research.