The goal of graph clustering is to partition a large graph into clusters based of node similarities and graph structure. Group of similar data types are done based on some measurements like distance between two vertices, weight of the node, etc. Here service oriented architecture is used in web service to obtain data from different web pages effectively and efficiently. The obtained data will be used to cluster the data based on attribute similarity. The clustered final graph is used to produce a static study about organization. This cluster will help to identify different communication and member’s interaction between the community members. This method will help to identify a community in social network or a group of members. In this paper, we are using modified similarity based graph clustering algorithm based on distance or difference between two web pages and weight of the node to group the data set.