Electronic Thesis and Dissertation Repository

Degree

Master of Science

Program

Geography

Supervisor

Micha Pazner

Abstract

Spatial databases are growing in size and complexity, yet current visual data mining methods are challenged when it comes to multivariate spatial data. The specific research question addressed in this thesis is: how can spatial multivariate data be effectively visualized using an icon based non-fused co-visualization approach? The thesis presents a Python based design and implementation of a visualization program termed GeoIcon Viewer. The program incorporates two different visualization methods: GeoIcon Image Map and Region-of-Interest Image Layers Chart. The GeoIcon Image Map technique uses an icon to co-visualize up to nine attributes at a single location. The Region-of-Interest Image Layers Chart method uses a small multiples approach to support the GeoIcon Image Map technique for data with negligible value differences. The thesis demonstrates the successful implementation of the GeoIcon Viewer with a case study involving remote sensing digital image analysis of a copper deposit. With the two visualization methods and eight input attributes, the GeoIcon Viewer generated real time interactive visualization outputs that can aid a user in multivariate spatial data mining.


Share

COinS