A Structural Approach to Contact Recommendations

People generally form network ties with those similar to them. However, it is not always easy for users of social media sites to find people to connect with, decreasing the utility of the network for its users. This position paper looks at different ways we can make friend recommendations, or suggest users to follow on the microblogging site Twitter. We examine a number of ways in which similarity can be defined, and the implications of these differences for community-building. People may want different things for friend suggestions, which has implications for both future research and product design.