This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. The OpenScience dataset comprises of research data of Slovenian universities (http://openscience.si/). The dataset is available in RDF/XML serialization format (http://www.w3.org/TR/rdf11-primer/). Resources (thesis, publication...) are uniquely identified by their National Repository IDentification number (NRID). Their base URI is "http://openscience.si/rdf/resource/" followed by their NRID. Authors are identified by their internal identification number (ID). Their base URI is "http://openscience.si/rdf/author/" followed by their ID. To describe each resource, we use the Dublin Core Metadata Initiative Metadata Terms (DCTERMS, defined at http://dublincore.org/documents/dcmi-terms/). The following properties are used: http://purl.org/dc/terms/abstract - abstract, defined zero to multiple times, possibly in different languages http://purl.org/dc/terms/title - title (and possible subtitle), defined only once http://purl.org/dc/terms/alternative - alternative title (and possible subtitle), defined zero to multiple times, possibly in a different language than title http://purl.org/dc/terms/issued - year of publication, defined only once http://purl.org/dc/terms/date - year of publication, defined only once, using the info:eu-repo date type (https://wiki.surfnet.nl/display/standards/info-eu-repo#info-eu-repo-DateTypesandvalue) http://purl.org/dc/terms/language - metadata language, defined zero to two times, using either ISO 639-2/T or ISO 639-2/B standard http://purl.org/dc/terms/source - place of publication, defined only once http://purl.org/dc/terms/subject - topic of resource represented using keywords (in possibly different languages) and the UDC classification code, defined zero to multiple times http://purl.org/dc/terms/type - the nature or genre of the resource represented using the COBISS typology (in English and Slovenian language, http://home.izum.si/COBISS/bibliografije/Tipologija_eng.pdf), COMARC/B type (in English and Slovenian language, http://unimarc2meeting.bnportugal.pt/documentos/9_ANKA%20ROGINA.pdf) and possibly ePrints type http://purl.org/dc/terms/identifier - unambiguous resource identifiers, defined zero to multiple times, possibly ISSN, ISBN, DOI and URN, using the info:eu-repo Alternative Identifier (https://wiki.surfnet.nl/display/standards/info-eu-repo#info-eu-repo-AlternativeIdentifier) http://purl.org/dc/terms/creator - author or contributer names and surnames, can be defined multiple times Note that not all properties will be defined for every resource, as some data may be missing or is otherwise unavailable. Furthermore, some data may be incomplete, incomprehensible or even incorrect. The problem stems from the fact that the data was aggregated from different sources which use different systems for cataloguing works as well as possible human errors while cataloguing.