data.dcsα



The Web of Linked Data is now a widely used distributed source of information for applications and agents. It links together resources and entities by typing links using formal semantics.

The data.dcs project investigates methods and techniques to create Linked Data describing people, research groups and publications within the Department of Computer Science based at the University of Sheffield.

Information within the department's web site is provided within legacy HTML documents and RSS publication feeds. Such information must therefore be leveraged in a machine-readable form so that it can then be woven into the Web of Linked Data.


Contents

1. About the Data
2. Data Access
3. Publications
4. Contacts




1. About the Data

Data is provided using Semantic Web technologies including the Resource Description Framework (RDF) and various ontologies. Information about the department is broken up into three axis: people, groups and publications. People can belong to multiple groups and groups contain many people. People author publications within groups but may also coauthor papers with members of other groups or the same group. This provides dynamics to information within the department which is hard to interpet without an explicit structure. We use the FOAF ontology - defined by the foaf namespace - to respresent person and group concepts and the bibtex ontology - defined by the bib namespace - to semantically defined information about publications.


1.1. Research Groups


Research Groups are teams within the department that contain researchers and PhD students. A research group is expressed as an instance of foaf:Group. Members of the department can be members of one or more groups at the same time. Research groups take the URI http://data.dcs.shef.ac.uk/group/ + the group name (e.g. "nlp" for the Natural Language Processing Group).



1.2. People


People belong to at least one group and are expressed as instances of foaf:Person. A given person produces publications and appears on various web pages within the department. Such appearances are expressed a statement connecting the URI of the person with the page on which they appear using the foaf:page relation. Person details are provided in the dataset including the person's email address (hashed using sha1sum and expressed using foaf:mbox_sha1sum) and the homepage of the person (related to the person using foaf:homepage). People are assigned as members of a given research group using foaf:member. A given person takes the URI http://data.dcs.shef.ac.uk/person/ + the person name (e.g. "Matthew-Rowe").



1.3. Publications


Publications define conference papers, journal papers, workshop papers, articles, books and theses. A publication is defined as an instance of bib:Entry and is given its title (using bib:title), the year of publication (using bib:hasYear) and the title of the book in which it was published (using bib:hasBooktitle). A person is said to have published a piece of work using foaf:made and a publication is said to have an author using foaf:maker.



1.4. Producing a graph-space

Putting the three axis together we are provided with a rich interpretation of information within the department of computer science. It allows questions to be asked over the information structure and discover interesting facts about the department. Below is an example of an information graph associating a given person (identified by the URI http://data.dcs.shef.ac.uk/person/Matthew-Rowe) with his publications and the research group he belongs to.







2. Data Access

By converting legacy data locked away within HTML documents and RSS feeds in to a machine-readable form we are provided with information linked together using semantically typed relations. This allows the information space to be queried in order to acquire additional conclusions. We provide a Linked Dataset contains URIs which can be dereferenced, check out some of the URIs below and browse the data.


2.1 Browsing Linked Data


All resources within Data.dcs are provided using Linked Data principles. This ensures that when a URI is dereferenced, whether it is a person, a publication or a group, that there is something at the end of that URI to provide a meaningful description about the resource. Try out some of the following URIs (visualised using the OpenLink's URIBurner Service):

<http://data.dcs.shef.ac.uk/person/Matthew-Rowe>
<http://data.dcs.shef.ac.uk/person/Fabio-Ciravegna>
<http://data.dcs.shef.ac.uk/person/Yorick-Wilks>

<http://data.dcs.shef.ac.uk/group/oak>
<http://data.dcs.shef.ac.uk/group/nlp>
<http://data.dcs.shef.ac.uk/group/vt>
<http://data.dcs.shef.ac.uk/group/spandh>
<http://data.dcs.shef.ac.uk/group/ml>
<http://data.dcs.shef.ac.uk/group/compbio>

<http://data.dcs.shef.ac.uk/paper/4156>
<http://data.dcs.shef.ac.uk/paper/3866>



2.2 Data Dumps


The data is also available as a raw dump as RDF/XML format here





3. Publications





4. Contacts

Managed and produced by Matthew Rowe from the Organisations, Information and Knowledge Group