Statistical Method for Visualizing the Spatial Pattern of Point Attributes

This article discusses the visualization for analyzing the spatial pattern of point attributes. Spatial aggregation is an effective tool for deriving measures that are useful for visual analysis. For instance, we calculate the measures of individuals, such as the average income and the proportion of the population over 65 in each administrative unit and visualize the results as choropleth maps. However, visual analysis of spatially aggregated data requires caution when point density varies across space. In such cases, distinguishing statistically significant from insignificant results can be difficult, as some results are based on many data points and others on few. To overcome this problem, this article proposes a new statistical framework for visualizing the spatial patterns of point attributes. Extending the frameworks of Kulldorff and Sadahiro, we developed a new framework that covers a wide variety of measures of point attributes. We took two concrete numerical measures to test the validity of the proposed framework. The applications indicated the proposed framework’s effectiveness for resolving the above-mentioned problem. A strength of the proposed framework is that it covers a wide variety of measures of point attributes, while existing methods are applicable only to a narrow set of measures.