Statistical techniques used for initial data analysis, such those used for checking statistical distribution, linearity, multicollinearity, etc. Based on results of an ESDA, initial hypotheses and further statistical data exploration can be developed, for example, by carrying out correlation, regression, and factor analyses.
Published in Chapter:
Mining Spatial Patterns of Distribution of Uranium in Surface and Ground Waters in Ukraine
Michael Govorov (Vancouver Island University, Canada), Viktor Putrenko (National Technical University of Ukraine “Kyiv Polytechnic Institute”, Ukraine), and Gennady Gienko (University of Alaska Anchorage, USA)
Copyright: © 2017
|Pages: 27
DOI: 10.4018/978-1-5225-0937-0.ch021
Abstract
A variety of geovisualization and spatial statistical methods can reveal spatial patterns in the distribution of chemical elements in surface and groundwater, and also identify major factors which define those patterns. This chapter describes a combination of modeling techniques to enhance understanding of large-scale spatial distribution of uranium in groundwater in Ukraine, by linking spatial patterns of several indicators and predictors. Factor, correlation, and regression analysis, including their spatial implementations, were used to describe the impacts of several environmental variables on spatial distribution of uranium. Local factor analysis (or Geographically Weighted Factor Analysis, GWFA) was proposed to identify major environmental factors which define the distribution of uranium, and to discover and map their spatial relationships. The study resulted in a series of maps to help visualize and explore the relationships between uranium and several environmental indicators.