Visual Data Mining for Quantized Spatial Data
Amy Braverman, Brian Hannon Kahn · 2004
In previous papers we’ve shown how a well known data compression algorithm called Entropy-constrained Vector Quantization (ECVQ; [ 3 ]) can be modified to reduce the size and complexity of very large, satellite data sets. In this paper, we discuss how to visualize and understand the content of such reduced data sets. We developed a Java tool to facilitate this using simple multivariate visualization, and interactively performing further data reduction on user selected spatial subsets. This enables analysts to compare reduced representations of the data for different regions and varying spatial resolutions. The ultimate aim is to explain physically observed differences, trends, patterns and anomolies in the data.