Measuring Data Abstraction Quality in Multiresolution Visualizations

Abstract

Data abstraction techniques are widely used in multiresolution visualization systems to reduce visual clutter and facilitate analysis from overview to detail. However, analysts are usually unaware of how well the abstracted data represent the original dataset, which can impact the reliability of results gleaned from the abstractions. In this paper, we define two data abstraction quality measures for computing the degree to which the abstraction conveys the original dataset: the Histogram Difference Measure and the Nearest Neighbor Measure. They have been integrated within XmdvTool, a public-domain multiresolution visualization system for multivariate data analysis that supports sampling as well as clustering to simplify data. Several interactive operations are provided, including adjusting the data abstraction level, changing selected regions, and setting the acceptable data abstraction quality level. Conducting these operations, analysts can select an optimal data abstraction level. Also, analysts can compare different abstraction methods using the measures to see how well relative data density and outliers are maintained, and then select an abstraction method that meets the requirement of their analytic tasks.

Integrating Data and Quality Space Interactions in Exploratory Visualizations

“…We hope to explore the linking of quality and data space to structure space. In such an extended linkage framework, we can consider performing tasks related to the quality of the data abstraction [5], such as highlighting datapoints based on their data abstraction quality, or examining the quality of sampling or clustering in a selected data subset.…”

Surveying the complementary role of automatic data analysis and visualization in knowledge discovery

“…Automatic algorithms could be run on extracted patterns to help the user assess their quality once they are detected. To date, the only systems we are aware of where a similar idea has been implemented are [11][12], where respectively data abstraction quality is measured and progressive automatic refinement of visual clusters is performed.…”

Visual analytics techniques for large multi-attribute time series data

“…Shneiderman's [BAS05] interactive pattern search provides fast visualization methods to retrieve information from large time series. Ward [CW06] uses histogram and nearest neighbor methods to measure data abstraction quality in multi-resolution visualization. Users can use both methods to evaluate how well the abstracted dataset represents the original data set.…”