Statistical visualization for data exploration: a case study on Sydney Olympic Park.

Abstract

Due to the powerful graphic interface in modern computers, visualization techniques have become more and more popular for data exploration. As a preliminary investigation, visualization is a greatly useful tool to lead to further statistical analysis and modelling. It is likely that historical environmental data from a range of different studies may contain useful information that should be assessed for determining management actions. However, these historical data are collected from many different studies, which have different space and/or time scales making the visualization process complicated. In this paper we present several ways of visualizing historical data based on a study at Sydney Olympic Park. The tools include spatial coverage and variation for different spatial resolutions, temporal coverage and time series plots for different time scales, multi-panel scatterplots for a small number of variables and "one-to-all" scatterplots for a large number of variables.