Meanwhile, also on the open data front, CiteULike makes data dumps available. CiteULike is a social bibliography service aimed at academic researchers. The service provides tagging ala del.icio.us, but is more vertical in that it understands URL references into many journal publishing services. CiteULike also provides hooks to upload common bibliography formats or download your bookmarks into common bibliography tools.
The CiteULike data dumps capture posting activity, linkouts (resolution of internal ids to actual articles), and group membership. This might be an interesting “starter” dataset for folks looking to practice application of network analysis algorithms, information retrieval techniques, or information visualization.