Social media has emerged as a free source of public data. Despite the diversity in social media platforms, current social media analysis tools consider them as a similar entity. Hence, these tools lack the flexibility to interpret the diversity of social media platforms and its analysis requirements. In order to interpret the diversity, we propose a novel service oriented approach to visualize and analyze social media platforms. Firstly, we formalize social media as a service and classify its functional and non-functional properties. Secondly, we conduct experiments on the real-world datasets for a variety of topics. We empirically quantify the functional and non-functional properties of social media platforms. We also present future directions and research challenges associated with the service orientation of social media.