Using the Self-Organizing Map to Visualize and Explore Socio-Economic Development

The socio-economic situation of a country can be measured in a number of ways, often by looking at indicator values describing different aspects of the social and economic reality in the country in question. Socio-economic development can be measured and observed by comparing yearly indicator values. As a large number of indicators are often required to accurately assess the socioeconomic development of a country, the dimensions of the analysis quickly become difficult. Composite variables, such as the Gini-coefficient and the Human Development Index, are often used, but drill-down analysis and visualization of the results reached with these is demanding and may be very laborious. The realm of data mining offers a number of tools for dealing with these problems. This paper utilizes the self-organizing map (SOM), a two layer unsupervised neural network, to observe, compare, and visualize development in 25 transition economies. Using longitudinal (1998-2002), multidimensional (14 indicators) socio-economic data, the paper presents how transition economies can be positioned, and their development tracked, with the SOM. The data is from the World Bank Group’s World Development Indicators (WDI) on-line database. The paper discusses the advantages achieved by using the SOM over simply comparing yearly indicator values, in observing and visualizing socio-economic development. The purpose of this paper is to show how the SOM can be used to observe, compare, and visualize development in transition economies. The paper does not seek to propose a new “correct” model for measuring socio-economic development, but to show, with examples, that the SOM could be a helpful tool in intuitively visualizing and comparing multidimensional socio-economic development.