Visual Analytics System for Large-scale Data Exploration

Date

2019-12

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

Visualization aims to produce meaningful visual representation of data, which has been shown to be a powerful means to reveal information (e.g., structures and/or relations) hidden in large-scale and complex data where our knowledge of these data is sparse. From the visualization of these data sets, important details are often di cult to discern due to the occlusion caused by the excessive visual elements presented. In this thesis, I present a visual analytics system that offers a complete pipeline for processing and visualizing these large-scale, complex data, starting from data cleaning and analysis to visualization generation and user interactions. In particular, my system supports a level-of-detail exploration scheme, showing the overview of the data in the beginning with various visual representations, followed by the detail-on-demand exploration. This thesis provides detailed description on the design and implementation of my system. To demonstrate the utility and generality of my system, I customized and applied it for the discovery of abnormal patterns hidden in traffic control data. New data cleaning, analysis, and visualization techniques are developed in order to address these traffic control data. My system successful revealed the errors in the data, including missing and misplaced entries. It also helped identify different traffic situations at different street intersections.

Description

Keywords

Information, Visualization

Citation