Various methods and techniques are prevailing for processing data to extract useful knowledge; nevertheless they are accounted for business transaction databases. Much data exists in non-business domains also, where there is an ardent need of extracting knowledge. Knowledge is useful for understanding data concisely or comprehending the concepts in a precise manner. Relational characteristics in data complicate the representation, and can be eased through graph structures. As one of the most general forms of data representation, the graph easily represents entities, their attributes, and their relationships to other entities. Envisioning the scope of the problems related to graph data, a collection of representations in graph data mining leads to the development of reference framework. In this paper we propose an integrated analytical reference framework with a comprehensive study and empirical evaluation.