Data.dcs is a project intended to produce Linked Data describing the University of Sheffield’s Department of Computer Science. At present the department’s web site contains important legacy data describing people, publications and research groups. This data is distributed and is provided in heterogeneous formats (e.g. HTML documents, RSS feeds), making it hard for machines to make sense of such data and query it. This paper presents an approach to convert such legacy data from its current form into a machine-readable representation which is linked into theWeb of Linked Data. The approach describes the triplification of legacy data, coreference resolution and interlinking with external linked datasets.