The increasing popularity of online social networks (OSN) plays an important role in how people communicate among themselves, and how they consume and propagate online materials. Social Network unique features of openness and rapid information sharing have been influencing communities and individuals, thereby making them perfect medium for expressing opinion, call for action, or breaking news. This thesis presents an extensible approach to analyse activities on OSN through the use of graph construction. The study used Twitter as application environment and observed three types of activity on the platform; trends, participations and interactions. Trends are perceived as links between topics and places, and they were used to investigate public concerns and interests. This study used trends data to provide measures for temporal activities and geospatial spread of topics. Users tend to relate themselves to topics through their participation. This relationship was used to investigate communities of users and activities. These communities were clustered based on selected entities from user and post objects. Also, the study employed interaction relationships to conduct analysis on selected accounts and how they interact with their audience. It presents various measure for performance evaluation, such as latency and popularity of accounts. In addition, indirect engagements were also discovered and found to be linked to direct interactions.