Mining, Modeling, and Leveraging Multidimensional Web Metrics to Support Scholarly Communities

The significant proliferation of scholarly output and the emergence of multidisciplinary research areas are rendering the research environment increasingly complex. In addition, an increasing number of researchers are using academic social networks to discover and store scholarly content. The spread of scientific discourse and research activities across the web, especially on social media platforms, suggests that far-reaching changes are taking place in scholarly communication and the geography of science. This dissertation provides integrated techniques and methods designed to address the information overload problem facing scholarly environments and to enhance the research process. There are four main contributions in this dissertation. First, this study identifies, quantifies, and analyzes international researchers’ dynamic scholarly information behaviors, activities, and needs, especially after the emergence of social media platforms. The findings based on qualitative and quantitative analysis report new scholarly patterns and reveals differences between researchers according to academic status and discipline. Second, this study mines massive scholarly datasets, models diverse multidimensional non-traditional web-based indicators (altmetrics), and evaluates and predicts scholarly and societal impact at various levels. The results address some of the limitations of traditional citation-based metrics and broaden the understanding and utilization of altmetrics. Third, this study recommends scholarly venues semantically related to researchers’ current interests. The results provide important up-to-the-minute signals that represent a closer reflection of research interests than post-publication usage-based metrics. Finally, this study develops a new scholarly framework by supporting the construction of online scholarly communities and bibliographies through reputation-based social collaboration, through the introduction of a collaborative, self-promoting system for users to advance their participation through analysis of the quality, timeliness and quantity of contributions. The framework improves the precision and quality of social reference management systems. By analyzing and modeling digital footprints, this dissertation provides a basis for tracking and documenting the impact of scholarship using new models that are more akin to reading breaking news than to watching a historical documentary made several years after the events it describes