Human creativity in the data visualisation pipeline

Abstract

There are some aspects of visualisation that are uniquely human. This is because data visualisation is heavily influenced by the science of human visual perception, much of which emerged from the Gestalt School of Psychology. Humans can recognise why certain aspects of the data matter, are aware of background information and use intuition, purpose and storytelling when choosing what to visualise. Intuition and visual perception are used in an iterative manner to craft the final visualisation. Whilst computer algorithms can create visualisations from data in a brute force combinatorial manner, humans are still much better at quickly determining what context and which aspects of the data, at what granularity, will successfully highlight what the data represents. Visual analytics, that combines machine learning, graphic user-interfaces and human interaction, is a popular way of addressing the shortcomings of fully automated computer generated visualisations. This paper is part of a larger project that will be exploring the development of a non-interactive computational algorithm that enhances the process of computer produced visualisations by introducing criteria and techniques from the theories of computational creativity, which is sub-field within the artificial intelligence domain. One of the objectives of the larger project is to identify the parts of the visualisation pipeline and also to identify what aspects of visualisation generation process humans are better at than computers – specifically with respect to human creativity. This literature review aims to address these identification objectives by means of a critical review of the parts of the visualisation pipeline.