The standard of living involves a wealth of different aspects ranging from health and education to the quality of the environment. It is a very time-consuming task to acquire insight into a statistical table of indicators that represents the numerous relevant aspects. Such a task is then a precisely suitable application for exploratory data analysis methods.
In Publication 2 a total of 39 indicators describing different aspects of the standard of living were chosen from a World Development Indicator set [World Bank, 1992], and a display of the structures of welfare and poverty that the data set reveals was set up using the SOM. The dominant axis on the map was found to be correlated with the GNP per capita, which was not included into the teaching corpus, by displaying the distribution of the GNP per capita values on the groundwork formed of the organized map. The interpretation of the fine structures of welfare and poverty can be done by inspecting the distributions of the values of the original indicators on the map groundwork.
In this study the input data set was finite, and the components of the data items were scaled to have equal variance, since it would have been difficult to determine differences in their importance. In the study it was also demonstrated how the missing value problem can be treated.