A comprehensive parallel coordinates system for air pollution analysis

Air pollution is one of the most serious environmental threats today. The analysis of its causes is an important research topic and visualization tools serve as the key component in weather data analysis. The goal of this thesis is to build a comprehensive parallel coordinates visualization system for air pollution analysis. Parallel coordinates is a fundamental visualization technique in information visualization. In this thesis, three major issues of parallel coordinates are addressed: visual clutter, dimension ordering and problematic transition. In the proposed system, a novel geometry-based clustering algorithm is proposed to reduce clutter in parallel coordinates. Three graph-based analytic tools are developed on the parallel coordinates. A selection graph provides an interface for users to select regions of interest in a cluttered display. A relation graph organizes clusters structurally which is an informative way for users to study inter-cluster relationships. A weighted complete graph aids users to find the optimal axis order based on its structure. The problematic transition is alleviated by the proposed non-linear growth technique and morphing transition on common operations in parallel coordinates. An analysis is carried out to demonstrate the effectiveness of the proposed system using real Hong Kong weather data