This activity undertaken by the HPC-Lab is organized into two main tasks: i) efficient social network analysis of location-aware services; and ii) social network analysis by means of spectral methods. In both tasks, research activities will pursue efficient methods to exploit state-of-the-art parallel programming paradigms such as MapReduce (for cluster systems) and MultiCore (for the single computational machine). Within i), we will study new problems and find efficient solutions to those problems within the realm of recommendation systems. As an example, we will consider recommending point of interests to members of a social network in a touristic location. In this scenario, the social network is fundamental to provide personalized recommendations to the tourist. Within the realm of personalization tools, we will identify user profiling techniques aimed at optimally combining information about content and structure of a user’s social network. Many other application scenarios will be considered during the project. In particular, we will exploit our expertise in the efficient implementation of spectral methods for social network analysis such as PageRank, HITS, Center-Piece Subgraph, etc. The design and implementation of spectral methods on the modern computational systems is our second research task. In particular, distributed and cloud systems are becoming increasingly used in recent years, and these systems are based on the MapReduce paradigm. In task ii), we will investigate the implications of using these computational systems (and related programming paradigm) on the design and implementation of social network analysis algorithms.