Zoning Systems of the Italian Territory for the Study of Rural Development

The present paper deals with the application of symbolic data analysis to data derived from the 2000 Agricultural Census. Symbolic data analysis is a powerful tool to represent classes of statistical units that may be derived a priori through a specific assertion or definition, or through an aggregation algorithm. In this study, farms are the basic units (first order objects) and local labor systems (a set of contiguous municipalities) are the units of analysis (second order objects). The features of the local labor systems are expressed by modal variables, a specific type of symbolic variable. A modal variable for a first orderobject is represented by a qualitative variable or a quantitative variable coded into a finite number of categories; for a second order object a modal variable is represented by the frequency distribution of the individuals into the categories. The empirical analysis is developed on those modal variables that describe a number of characteristics of the farms within the local labor systems. Through a cluster analysis on modal variables, a partition of the Tuscan local labor systems has been obtained.