Space optimisation in multi-dimensional data visualisation

Publication Type:
Issue Date:
Full metadata record
Files in This Item:
Filename Description Size
01front.pdf432.5 kB
Adobe PDF
02whole.pdf9.98 MB
Adobe PDF
Multi-dimensional Data Visualisation (MDV) is the technique to generate visual presentations of datasets with more than three features (or attributes). These graphic representations of data and associated data features can facilitate human comprehension, extraction of implicit patterns and discovery of the relationships among numerous data items for visual data analysis. Although many optimisation methods have been proposed in the past to improve the visual data processing, not many have been applied to MDV. In particular, little research work has been done in the field of display space optimisation. This thesis focuses on the optimisation of two popular Multi-dimensional visualisation methods: 1) scatterplot matrix and 2) parallel coordinate plots, visualisation by using unique approaches to achieve display space optimisation and interaction. The first contribution of the thesis is proposing a new visualisation approach named the Spaced Optimised Scatterplot Matrices that achieves the maximization of the display space utilisation through position transferring. Breaking through the limitation of discovering the pairwise variable relationships, the new method is able to explore the influences of a single variable towards others. In addition, our algorithms improve the efficiency of interactive Multi-dimensional data visualisation significantly, through the reduction of the computational cost. The second contribution of the thesis is to improve the parallel coordinate plots and apply it to the computer forensic investigation. As we are living in a big data era, it is much harder for the researchers to provide accurate evidence for victims within a certain time frame. Our research shows that visualisation techniques can improve the working efficiency of investigations in certain cases. To conclude, we propose a concept of a space optimised Scatterplot Matrix(SPM) visualisation technique considering the shortcomings of the exsited SPM and parallel coordinates in Multi-dimensional visualisation research area. In the meantime, to demonstrate the necessity of our research methodologies, we apply them into computer forensics, which is an area needed analyzing abilities with higher accuracies and efficiencies. By the tests on using Parallel Coordinates and DOITrees, the forensic specialists can easily discover the necessary information in different cases. In the future, we plan to improve our space optimised scatterplot matrix technique from technological optimisation and the broaden application aspects. For example, dealing with the visualisation coolusion problem, non-trivial computation time issue, etc; We will also do the investigations to enlarge the development and availability of Multi-dimensional visualisation techniques.
Please use this identifier to cite or link to this item: