RNTI

MODULAD
Analyzing European Social Survey data using symbolic data methods and Syrokko software
In SDAV 2015, vol. RNTI-E-29, pp.89-100
Résumé
The paper presents an application of Symbolic Data Analysis (SDA) with SYR software of Syrokko company to the fifth round of the European Social Survey (ESS) carried out among European inhabitants. In this study, we are not interested in studying the people themselves but by the comparison of different European countries, or different regions of Europe (Western Europe, Eastern, Nothern, Southern), or some groups of inhabitants by age, gender, region, etc. Here, we study, however, the 52 European countries by age groups. We describe each of them by all the results of its inhabitants using symbolic data. Symbolic Data Analysis (SDA) is proving so useful to aggregate up micro data (at the level of the inhabitants) to higher level units called concepts (the countries or European regions, for instance), using symbolic barchart or interval-valued variables. Using this aggregation we lose less information than occurs in classic analysis because symbolic data allow keeping the variation within the concepts. It allows keeping the variation of the results at the level of the inhabitants when they are aggregated up to their country. Hence, this could be called as “smart aggregation”.