Enabling Ubiquitous OLAP Analyses

An OLAP analysis session is carried out as a sequence of OLAP operations applied to multidimensional cubes. At each step of a session, an operation is applied to the result of the previous step in an incremental fashion. Due to its simplicity and flexibility, OLAP is the most adopted paradigm used to explore the data stored in data warehouses. With the goal of expanding the fruition of OLAP analyses, in this thesis we touch several critical topics. We first present our contributions to deal with data extractions from service-oriented sources, which are nowadays used to provide access to many databases and analytic platforms. By addressing data extraction from these sources we make a step towards the integration of external databases into the data warehouse, thus providing richer data that can be analyzed through OLAP sessions. The second topic that we study is that of visualization of multidimensional data, which we exploit to enable OLAP on devices with limited screen and bandwidth capabilities (i.e., mobile devices). Finally, we propose solutions to obtain multidimensional schemata from unconventional sources (e.g., sensor networks), which are crucial to perform multidimensional analyses.