As we know, people are connected in various social networks such as phone calls, emails and Facebook. Then an interesting question arises whether we can improve our chance of making friends with some particular stranger in the network? Here the stranger can be our potential spouse, colleague, advisor or boss. Of course we can directly make a phone call, send an email or send an invitation through Facebook to the stranger. But we are very likely to be ignored or reported since we are completely unknown to the stranger. Moreover, private information such as email addresses and phone numbers are unavailable most of time.

One straightforward solution is to link one by one along the shortest path to the target. However, the problem becomes complicated when the shortest path is still long or when there exists more than one shortest path, especially in a dense network. In this report, I develop a strategy based on network topology including shortest paths, community structure and eigenvector centrality to deal with these problems. The method and the accompanying analysis will be illustrated on a dataset covering email communication of about 150 users of Enron, an American energy company.