A Reference Architecture for Social Media Intelligence Applications in the Cloud

As the social media upsurge of today continues to mount, opportunities to derive collective intelligence from online social networking (OSN) content sources are inevitably expected to grow. While enterprise organizations and research institutions make a dash for identifying rich insights and opportunities to tap into the millions of conversations and user profile relationships exposed by this new social-influenced big data phenomenon, architectural concerns regarding the storage and processing of large datasets unearthed by OSNs, along with performance, scalability, fault-tolerance, security, privacy, and high-availability solutions have become an area of concern for social media intelligence (SMI) solutions. In this literature, we present a reference architecture, for designing SMI solutions. In addition, we showcase two key case studies for SMI applications built on this architecture. Our selected case studies are focused on the analysis of User-Generated Content (i.e. With Sentiment Analysis in Twitter data) and Social Graph Influence (i.e. In a Facebook-influenced Movie Recommendations solution). We evaluate the ´goodness-of-fit´ in applying our model to these case study solutions and present results from our performance evaluation of these cloud-hosted solutions across multiple cloud providers like Amazon AWS, Microsoft Azure and Google Cloud.