Cluster analysis has become a popular instrument in determining the innovativeness and competitive power of national and regional economies. A wide variety of methods have been employed by researchers to identify the key industrial clusters in a regional economy. This paper reports the results of practicable methodologies for identifying aggregate economic clusters so-called megaclusters in Andalusia, using the last available input output data (year 2000). The first method used, based in graph theory, classifies sectors according to substantial mutual dependency through their deliveries and purchases. The second one, based on the application of principal components to the input-output table, provide clusters according to the similarity of the intermediate purchases and/or the intermediate sales structures.