Towards Smart Geovisualization based on the Data-driven Cube Concept

Abstract

With the breakthrough of information technology social media and sensing technology a large amount of data has accumulated rapidly in a random or organized way "Big Data" has received tremendous attention in recent years People use many innovative technologies such as artificial intelligence and deep learning to explore new knowledge or new information hidden in the increasingly huge volume of data Visualization is one of the steps of big data analysis The data in real world has two major consideration "space" and "time" How to highlight these two aspects in the development of geographic information design and processing so as to explore the smart applications based geographic information will be an important challenge to the future research The research purposes a cube-based concept of geographic data organization Three major characteristics of geographic data namely temporal spatial and variables are selected as the three major modelling perspective of the cube Each axis has different details based on the numbers of attributes or the complexity of their respective data recording methods Each geographic data can be classified and modelled by a single cube element based on its distinguished characteristics According to the hierarchical design cube elements can be consolidated or disassemble to generate new information following specified rules of data quality As every cube element is defined with a unique set of distinguished geographic characteristic it has specific assumption requirement and corresponding visualization methods After summarizing various prerequisites of visualization methods this research suggests the appropriate visualization methods for each cube element through two-way analysis To ensure the correct definition and classification of geographic data characteristics data combination and visualization methods this research also proposes quality constraints for the processing of geographic data This can smartly select the appropriate visualization strategies and ensure the correctness of visualized results The recent trend undoubtedly indicates there is a high application value hidden in the big data However the lack of effective management analysis and visualization strategies will cause heavy and unnecessary processing work It seriously hinders the application of big data This research focuses on the effective induction of the characteristics of geographic data using management and knowledge mechanisms to explore hidden information and finally displaying the information with appropriate visualization techniques to decision makers The advantage of this intelligent mechanism is to bridge the gap between decision-makers and the huge volume of unfamiliar domain data so as to achieve the goal of cross-domain and time series data interoperability for enabling flexible and multi-purpose applications While the current visualization software mainly focuses on simplifying the operation of functions and seldomly focuses on the combination of geographic data knowledge automatic analysis and judgment this research provides a solid foundation for the effective combination of visualization software and big data database and have the potential to drive the evolution of geographic information software

Date of Award	2021
Original language	English
Supervisor	Jung-Hong Hong (Supervisor)