Multiple view systems are often used by visualisation developers. They are useful for displaying or interpreting data in several, multiple or parallel ways. One of the reasons developers use such duplication is to help clarify the information. Perhaps a user understands one style of visualisation better than another, or perhaps one type of visualisation form makes it easier to perform a particular task, whereas another form makes it easier to perform a different task. For these purposes, there are many visualisation tools and programming libraries that help users create multiple view visualisations. However, it is not easy for a new developer to know how to lay out and position the views in their systems, how many views they should use, what is the best visualisation type for each view, or what design attributes work best. Therefore, developers and learners should have guidelines and frameworks to assist them in making the right design decisions.
The long-term vision of this research is to develop theories for data visualisation, and develop specific guidelines on best practices for multiple view systems; this will assist researchers and developers in understanding and developing multiple view systems. To achieve these goals, guidelines need to be based on current practice. Our methodology is to perform an in-depth quantitative investigation to understand best practices for what researchers currently do with multiple view systems. From this investigation guidelines can be developed. Furthermore, such guidelines could be programmed into a grammar, and into a multiple view design tool, which would help developers create multiple view visualisations.
Therefore, this work focuses on investigating multiple view systems, in order to develop a set of guidelines and a system to help new developers to create multiple view visualisations and make the correct design decisions. For this purpose, we designed a visualisation tool based on a quantitative evaluation of multiple view systems. This grammar-based tool allows learners to create, control and save multiple view visualisations in a simple way by using a multiple view grammar.
The research focuses on eight research questions that are used to structure this thesis, ranging from counting views, quantifying views, developing a multiple view grammar, and creating a multiple view tool: (1) What are the strategies for selecting which multiple view images to evaluate? (2) What are the strategies for defining and determining a “view” in multiple view visualisation? (3) What are the strategies for coding multiple view topologies and visualisation types? (4) How many views should developers use in multiple view systems? (5) What layout arrangements are popular in multiple view systems? (6) What visualisation types are used in each view and what types of visualisation come together? (7) What salient guidelines can be learnt from the analysis, to assist users in developing multiple view visualisations? (8) What is a multiple view grammar and how is it used to create a multiple view layout?
By tackling and answering these research questions, the thesis makes six novel research contributions. First, it introduces a strategy for selecting which images of multiple view visualisations to evaluate (Chapter Three). Second, this research creates a strategy to help researchers ascertain “what constitutes a view” in a multiple view visualisation (Chapter Three). Third, through statistical analyses of multiple view visualisations, this research produces results of a comprehensive quantitative analysis of multiple view visualisations, which can help researchers to conduct further investigation on multiple view systems (Chapter Four). Fourth, from this analysis, the thesis develops a set of guidelines to help novices in the data visualisation field, as well as developers, to create robust multiple view visualisations (Chapter Four). Fifth, this research introduces a new grammar to create multiple view layouts by using the concept of cutting a view vertically or horizontally to create two views (Chapter Five). Sixth, this work develops the LayMV tool to create, control and save multiple view visualisation, based on the analysis of the multiple view systems and the set of guidelines for creating multiple view systems. The LayMV tool uses the multiple view grammar to create and manage multiple view visualisations (Chapter Six).
In conclusion, this dissertation provides learners and practitioners with an in-depth analysis of the multiple view field, which can help them create multiple view visualisations and carry out further investigations on multiple view systems. In addition, the LayMV tool and the multiple view grammar can help users to create, control, save and reload multiple view systems.