Social Web describes a new culture of participation on the Web where more and more people actively participate in publishing and organizing Web content. As part of this cul- ture, people leave a variety of traces when interacting with (other people via) Social Web systems. In this paper, we investigate user modeling strategies for inferring personal interest proles from Social Web interactions. In particular, we analyze individual micro-blogging activities on Twitter. We compare dierent strategies for creating user proles based on the Twitter messages a user has published and study how these proles change over time. Moreover, we evaluate the quality of the user modeling strategies in the context of personalized recommender systems and show that those strategies which consider the temporal dynamics of the individual proles allow for the best performance.