Immersive Multidimensional Data Visualisation using Geon Based Objects
MetadataShow full item record
In an ever-expanding technologically driven society large amounts of data are being generated daily. The sheer volume of data is becoming a challenge to understand or use in the decision-making process. This is not only limited to businesses, politicians or scientists but it is becoming more relevant to the general public as well. The data has become a commodity, but its value is closely tied to our ability, as data owners, to extract meaning from it. Typically, the process is complex and mainly reserved for proficient data scientists who are familiar with the process of extracting information and the specialised tools used for that process. With recent technological advancements in Virtual Reality technology, especially with a new range of affordable hardware such as Head Mounted Displays (HMD), visual analytics fused into a new research field called Immersive Analytics. Its main purpose is to focus on analytical reasoning with the help of immersive computer interfaces and to explore the human ability to perceive and interact with these representations as being real objects. This thesis investigates multi-dimensional data representation using simple 3D geometric shapes called Geons that form the basis for a complex visual object (glyph) presented to the user in an immersive virtual environment. A set of rules, for building the glyph, was created from the basic principles of object recognition theory called Recognition by Components. A toolkit was created that is capable to represent multi-dimensional data sets in immersive virtual environments and it has at its core a human-centred approach. Careful consideration was given to the immersive aspect of the application with a focus on spatial immersion, data embodiment, multi-sensory presentation, and immersive narrative. A series of experiments was carried out to evaluate the effectiveness of this approach including the evaluation of the immersive aspect of the experience. The results indicate that applying theories of structural object recognition to the construction of complex visual objects can facilitate the identification of optimal solutions in large data sets without the user having any prior experience in data exploration. The findings also show that the immersive aspect of the application has a significant contribution to the sense-making process and the participants reported positive feedback in measuring the levels of immersion.