Mendeley's open data for science and learning: a reply to the DataTEL challenge. (1st August 2012)

• Record Type: Journal Article
• Authors: Jack, Kris; Hristakeva, Maya; Garcia de Zuniga, Rosario; Granitzer, Michael
• Is Part Of: International journal of technology enhanced learning. Volume 4:Number 1/2(2012)
• Page Start: 31
• Page End: 46
• Publication Date: 2012-08-01
• Subjects: digital libraries -- recommender systems -- data sets -- DataTEL -- Mendeley; Educational technology -- Periodicals; Education -- Effect of technological innovations on -- Periodicals; Web-based instruction -- Periodicals; Computer-assisted instruction -- Periodicals; 371.3305
• Languages: English
• ISSNs: 1753-5255

Abstract

In this paper two open access data sources, provided by Mendeley Ltd., are described in detail. The first data source is a snapshot of 50, 000 user libraries, enabling researchers to test collaborative filtering algorithms for scientific paper recommendation. The second data source is Mendeley's API that provides access to 50 million research articles, crowdsourced from over 1.5 million users. This API allows researchers to build third party applications on top of real-time and large scale data. By providing data sources for the study of collaborative filtering, metadata extraction, deduplication and related research, Mendeley hopes to enable researchers in these domains to better collaborate and share knowledge, ultimately encouraging good research practices and advancing the state of scientific knowledge.