The access information law approved by the Brazilian government regulates the provision of open government data in the Web. However, they are heterogeneous, unstructured and derived from independent sources, making it difficult to interconnect. This paper presents a process of identifying sources, ontology generation, mapping and publishing statistical linked data in the form of multidimensional cubes, represented by the RDF Data Cube Vocabulary. In this process, data are transformed and assigned semantic meaning through its connection with domain ontologies. Through a web application, the publication of these data is automated, allowing for future analysis operations with the use of On-Line Analytical Processing (OLAP).