Towards synthetic social media data

Abstract

There is an increasing need for training and testing data that can be used for the development of technologies and research. Due to data protection regulations, a lot of the real-world data – especially the data from social media – becomes unavailable to use. The problem can be solved by generating synthetic data that imitates the properties of real-world data. In this paper, we present Fabulator – a social media generator that combines text and graph structure to be used for the generation of synthetic data and, in the future, for the simulation of various events on social media.