https://hdl.handle.net/10388/14475

Wang, Shisong (2023) Set-Stat-Map: Visualizing Spatial Data with Mixed Numeric and Categorical Attributes, USASK Harvest - Theses and Dissertations, https://hdl.handle.net/10388/14475

Multi-attribute datasets are common and appear in many important scenarios for data analytics. Such data can be complex and thus difficult to understand directly without using visualization techniques. Existing visualizations for multi-attribute datasets are often designed based on attribute types, i.e., whether the attributes are categorical or numerical. Parallel Coordinates and Parallel Sets are two well-known techniques to visualize numerical and categorical data, respectively. However, visualization for mixed data types appears to be challenging. A common strategy to visualize mixed data is to use multiple information-linked views, e.g., Parallel Coordinates are often augmented with maps to explore spatial data with numeric attributes. In this paper, we design visualizations for mixed data types, where the dataset may include numerical, categorical, and spatial attributes. The proposed solution Set-Stat-Map is a harmonious combination of three interactive components: Parallel Sets (visualizes sets determined by the combination of categories or numeric ranges), statistics columns (visualizes numerical summaries of the sets), and a dataset-specified map view (geospatial map view for spatial information, heatmap for pairwise information, etc.). We also augment the Parallel Sets view in two main ways: First, we impose textures on top of colors, which are spread into the other views, to enhance users' capability of analyzing distributions of pairs of attribute combinations. Second, we limit the number of sets for each axis to a small number by merging some of them into one and limit the sizes of the merged sets to improve the rendering performance as well as to reduce users' cognitive loads. We demonstrate the use of Set-Stat-Map using different types of datasets: a meteorological dataset (CFSR), an online vacation rental dataset (Airbnb), and a software developer community dataset (StackOverflow). We provide design guidelines based on the results of the analysis of the performance from both visual analytics and scalability aspects. To examine the usability of the system, we collaborated with meteorologists, which reveals both challenges and opportunities for Set-Stat-Map to be used for real-life visual analytics.

https://harvest.usask.ca/bitstreams/d832399e-d017-4ca6-ae9e-68e1caa32cf2/download