Outlier highlighting for spatio-temporal data visualization
Ulla Pyysalo, Juha Oksanen · Cartography and Geographic Information Science · 2013
This paper compares several different visual highlighting methods that are applied to spatio-temporal data to support analytical reasoning based on situation awareness maps. In our work, visual highlighting is treated as a method for visual analytics intended to draw visual attention to particular observations in a map interface without mouseover or other forms of interaction by the user. The case study uses air traffic observations around the Baltic Sea for a 12 hour period of time. These observations are illustrated on the background map in three layers, namely, the base map layer, the thematic raster layer, and the symbol layer. The map design was chosen to emphasize the visual hierarchy of these layers. Some of the air traffic observations were outliers because the flights were in an unexpected location or the speed of traffic was exceptional. In the visualization experiments, outliers were flagged in the hierarchy via highlighting in order to study how well different highlighting methods communicate their quality and quantity. The comparison of methods was carried out by setting up a user survey. The paper summarizes the answers to the questionnaire and analyzes and discusses some highlighting challenges.