A Data-Driven Approach to Categorize Climatic Microenvironments

Abstract

In urban climatology, identifying areas of similar microclimatic conditions helps to relate fine-scale urban morphology variations to their impact on atmospheric surroundings. Mobile transect measurements yield high-resolution microclimate data that allow for the delineation of these areas at a fine scale. However, the resulting spatio-temporal multivariate data is complicated and requires careful analysis and visualization to identify the emergent climatic microenvironments. Our previous work used a glyph-based visualization to comprehensively visualize spatially aggregated multivariate data from mobile measurements over diverse routes. This aggregation was conducted over a regular grid, and the utilized glyphs encoded multivariate relationships, average wind direction during data collection, number of transects traversing a grid cell, and grid cell size. In this paper, we reduce the visual complexity of the resulting map by coloring the background of the grid cells based on a comparison of the glyphs. The result is a gridded map that visually emphasizes spatial zones of similar multivariate relationships and that takes the information encoded by the glyphs into account. A preliminary evaluation shows that the described approach yields zones that line up with the physical structure of the study site.