This paper addresses the opportunities and challenges of the use of open data as input for research that would be disseminated through Open Access journals. In particular, it seeks to determine the technical and legal requirements that data must meet to be truly open, under the premise that only in this way will likely complete the cycle of open research and publication.

The open data movement, based on the opening and reuse of information in the science and the public sector worldwide, has been greatly enhanced in recent years thanks to the development of information and communication technologies. Not only because of their contribution to facilitate and enhance the access and exchange of information, but especially, for the evolution in their capabilities for the analysis, exploitation and processing of data. They are making possible to build collaborative knowledge from data extracted and released from multiple sources, to verify the research results and reproduce scientific experiments, as well as the creation of services and applications based on the use, aggregation or combination of data.

There are certain three fundamental conditions that data must satisfy in order to be open, and therefore, allowing reuse: i) socially open, as they have to be published on the web, allowing free and open access, ii) legally open, thus explicitly licensed in a way that permits commercial and non-commercial use and re-use without restrictions, just requiring source attribution and citation to potential re-users, and iii) technically open, therefore available in a machine-readable format along with their metadata, so as to ensure their interoperability. These principles are initially shared with the Linked Open Data (LOD) principles, as defined by Tim Berners-Lee in his five-star incremental scheme.

Several national governments and many other institutions have created web sites to distribute their “open data”. However, at this stage, no criteria and mechanisms have been established yet to ensure that these data sets have actually been published in open formats and licenses, even less according to other LOD principles.

It is necessary to establish the minimal requirements for truly open data, both in its legal and technological dimensions, to ensure the effective and efficient exchange of data between producers / authors and re-users / researchers, and help raising the level of reuse and citation.

With this aim, we propose the application of a matrix to evaluate the degree of openness of data, data sets and databases, structured around two main axis: on the one hand, the license they hold, ranked according to the permissions, limits and restrictions imposed, and, on the other hand, the file format, following the ISO 2146/2010 classification, together with the Tim Berners Lee five stars scale. We propose an equation for determining the degree of data openness based on the balanced combination of both values.

Some preliminary results are presented, after having applied the matrix to a selection of public data potentially useful as raw research data in Social Sciences. The validity of this formula and the feasibility of its application in open data portals, are discussed.