Categorising data visualisations: a study of figurative images' impact on cognition and engagement and their production

As data is becoming increasingly ubiquitous and relevant to our daily lives, there has been a growing need to communicate data-driven information to a broad audience. Despite abstract and minimalist data graphs are widely used, they are not equally useful for all individuals. It still remains challenging for the “ordinary” non- "data expert" to comprehend data visualisations that use only abstract geometric shapes. This thesis therefore specifically investigated the potential benefits of using particular types of figurative images in data visualisation, especially their impact on cognition. Firstly, a literature review was conducted to understand how other researchers had examined the use of figurative images in data visualisation and key related cognitive theories such as semiotics, multimedia learning, and metaphor to establish the theoretical foundation for the primary research conducted. Then, an image-based reclassification of data visualisation was undertaken to more fully understand the differences in shapes/images used in constructing visual representations of data. The findings presented a more nuanced approach to classifying data visualisation that further extends the traditional binary classification system used in previous studies (comparing "plain" data graphics to "embellished" data graphics). A novel VE (Visualisation Expression) categorisation system consisting of four VE types (Abstract VE, Literal VE, Context VE, and Metaphorical VE) was established for use in data visualisation. Furthermore, the researcher compared the effectiveness of data visualisation across four VE types through a set of rigorous experiments. The empirical analysis suggested that, when used appropriately, the use of figurative images to construct visual representations of data can provide meaningful improvements in viewers' understanding, memory, and engagement with data-driven information compared to abstract geometric shapes. Finally, a design workshop was run to examine the usefulness of the newly constructed VE categorisation approach in guiding the creation of varied figurative data visualisations. By integrating student designers' reflections on their creative process, a design process model was formulated, which further provides more specific creative steps based on the definitions of four VE types. This was done in order to more effectively guide designers and others involved in presenting data to produce more understandable, memorable, and engaging data visualisation using particular types of figurative images. These findings are an important contribution to knowledge and will inform the practice and research of figurative data visualisation in the future.