Exploring spatial data with dominant attribute map and parallel coordinates

This paper suggests a technique and a corresponding software tool for exploratory analysis of multivariate numeric data associated with area geographical objects such as units of administrative division of a territory. More specifically, the paper considers analysis of several comparable attributes, i.e. those measured in the same units and semantically related. Examples of comparable attributes can be provided by attributes representing proportions in total (such as division of population by age group or land use statistics) or rates (such as birth and death rates). The technique suggested is based on the use of a parallel coordinate plot dynamically linked to an interactive map by means of simultaneous highlighting of corresponding objects. The map represents by painting classification of the geographical objects according to the dominant attribute: the colour of an object corresponds to the dominant attribute while the degree of darkness shows the strength of the dominance. The tool described enables various transformations of the parallel coordinate plot. The transformations help the user to consider data from different perspectives. Any transformation is immediately reflected by the map. The map and the parallel coordinate plot mutually facilitate the interpretation and enhance analysis-supporting capabilities of each other.