A vast increase in computing power has enabled Social Network Analysis (SNA) to find its application in a variety of disciplines beyond its traditional use as a tool for sociologists to understand human networks. A few researchers in marketing has begun to be interested in applying SNA to investigate network structures formed of co- purchase relationships among products in the market. However, prior research only stopped at examining overall characteristics of the network, limiting its analytical capability to limited evaluation criteria such as density and degree contribution of the entire network. This research extends the SNA-based co-purchase analysis framework by incorporating the inter-community and intra-community characteristics of the network. We tested our framework by applying it to the analysis of annual ID-POS data on 36,981 customers of a supermarket in Fukuoka, Japan. Results demonstrated the capability of the framework to identify relationships among product communities and intra-community roles of products.