Statistical Data are considered one of the major sources of information and are essential in many fields as they can work as social and economic indicators. A statistical data set comprises a colletion of observations made at some points of a logical space and is often organized as what is called a data cube. The proper definition of the data cubes, especially of theis dimensions, helps processing the observations and, more importantly, helps combining observations from different data cubes. In this contexto, the Linked Data principles can be profitably applied to the definition of data cubes, in the sense that the principles offer a strategy to provide the missing semantics of the dimensions, including their values. This dissertion first describes a mediation architecture to help describing and consuming statistical data, exposed as RDFtriples, but stored in relational databases. One of the features of this architesture is the Catalogue of Linked Data Cube Descriptions, which is described in detail in the dissertation. This catalogue has a standardized description in RDF of each data cube actually stored in statistical (relational) databases. Therefore, the main discussion in this dissertation is how to represent the data cubes in RDF, i.e., how to map the database concepts to RDF in a way that makes it easy to query, analyze and reuse statistical data in the RDF format.