MESUR: Usage-based metrics of scholarly impact

The evaluation of scholarly communication items is now largely a matter of expert opinion or metrics derived from citation data. Both approaches can fail to take into account the myriad of factors that shape scholarly impact. Usage data has emerged as a promising complement to existing methods of assessment but the formal groundwork to reli- ably and validly apply usage-based metrics of scholarly im- pact is lacking. The Andrew W. Mellon Foundation funded MESUR project constitutes a systematic eort to define, validate and cross-validate a range of usage-based metrics of scholarly impact by creating a semantic model of the scholarly communication process. The constructed model will serve as the basis of a creating a large-scale semantic network that seamlessly relates citation, bibliographic and usage data from a variety of sources. A subsequent program that uses the established semantic network as a reference data set will determine the characteristics and semantics of a variety of usage-based metrics of scholarly impact. This paper outlines the architecture and methodology adopted by the MESUR project and its future direction.