Andrea Giommi, Dr., Alessandra Petrucci, Ph.D., and Christian T. Brownlees, Ph.D. Department of Statistics, University of Firenze, Viale Morgagni 59, Firenze, 50134, Italy
In the last years, the interest in the territorial dimension of the economic development has considerably strengthened. In the meantime, new methods and technologies including the Geographical Information Systems (GISs), that allow to relate statistical data coming from different and heterogeneous sources to areas or dots of a surface, have been developed and become widely available.
The results of the most of current surveys in Italy are often referred to wide territorial areas, and this level of detail in the information is not always adequate to fulfill the requirements necessary for private and public operators to face the new challenges of the economy. In order to reduce the information gap and get adequate territorial reference for decision support it is necessary to integrate sampling and census data with register and administrative data and define a suitable zoning of urban and rural areas which can be use for both analysis and data dissemination and as a basis for redesigning surveys.
In a previous IAES conference we outlined some methodological proposals for defining homogeneous geographical zones at a level sufficiently detailed to be useful for agricultural policies. In this paper we advance a new methodology called MOSCA (Multidirectional Optimum Spatial Clustering Algorithm) which can easily been applied to both urban and rural areas in order to detect anomalous zones and/or to cluster spatial units in the presence of territorial constraints.