As a representative Web 2.0 application, collaborative tagging has been widely adopted and inspires significant interest from academies. Roughly, two lines of research have been pursued: (a) studying the structure of tags, and (b) using tag to promote Web search. However, both of them remain preliminary. Research reported in this paper is aimed at addressing some of these research gaps. First, we apply complex network theory to analyze various structural properties of collaborative tagging activities to gain a detailed understanding of user tagging behavior and also try to capture the mechanism that can help explain such tagging behavior. Second, we conduct a preliminary computational study to utilize tagging information to help improve the quality of Web page recommendation. The results indicate that under the user-based recommendation framework, tags can be fruitfully exploited as they facilitate better user similarity calculation and help reduce sparsity related to past user-Web page interactions.