Scholarly publishing has seen an ever increasing interest in Linked Open Data (LOD). However, most of the existing datasets are designed as flat translation of legacy data sources into RDF. Although that is a crucial step to address, a lot of useful information is not expressed in RDF, and humans are still required to infer relevant knowledge by reading and making sense of texts. Examples are the reasons why authors cite other papers, the rhetorical structure of scientific discourse, bibliometric measures, provenance information, and so on. In this paper we introduce the Semantic Lancet Project, whose goal is to make available a LOD which includes the formalisation of some useful knowledge hidden within the textual content of papers. We have developed a toolchain for reengineering and enhancing data extracted from some publisher's legacy repositories. Finally, we show how these data are immediately useful to help humans to address relevant tasks, such as data browsing, expert finding, related works finding, and identification of data inconsistencies.
