Quantcast

An IconMap-based exploratory analytical approach for multivariate geospatial data

Research paper by XianFeng Zhang, ChunHua Liao, Yu Liu, Jonathan Li

Indexed on: 20 Oct '12Published on: 20 Oct '12Published in: Science China Information Sciences



Abstract

This paper discusses an iconmap-based visualization technique that enables multiple geospatial variables to be illustrated in a single GIS raster layer. This is achieved by extending the conventional pixel-based data structure to an iconic design. In this way, spatial patterns generated by the interaction of geographic variables can be disclosed, and geospatial information mining can be readily achieved. As a case study, a visual analysis of soil organic matter and soil nutrients for Shuangliu County in the city of Chengdu, China, was undertaken using the prototype IconMapper software, developed by the authors. The results show that the static IconMap can accurately exhibit trends in the distribution of organic matter and nutrients in soil. The dynamic iconmap can both reflect interaction patterns between organic matter and the nutrient variables, and display soil fertility levels in a comprehensive way. Thus, the iconmap-based visualization approach is shown to be a non-fused, exploratory analytical approach for multivariate data analysis and as a result is a valuable method for visually analyzing soil fertility conditions.