Challenges and Opportunities for detecting and measuring diffusion of scientific impact across heterogeneous altmetric sources

Alternative Metrics (also known as altmetrics) ­ alternative measures of scholarly impact has achieved increasing recognition among scientists and researchers of scientometrics, the field dedicated to measuring science, technology and innovation. Currently our lab is working on developing an innovative platform for incorporating diverse sources of scientometric data; both traditional (e.g. bibliometric) and new (e.g. social media) in order to capture a comprehensive view of scientific practice and discourse as well as the diffusion of scientific impact across the Web and social media ­ a type of altmetrics analytics dashboard. We argue for an approach based on ingesting and integrating vast amounts of social media and traditional mainstream web content into a large knowledge graph for analysis. Examples of data sources include such as mainstream news, blogs, microblogs, data from official government agencies i.e. white papers, legislation and policies and of course traditional bibliometric sources. This involves challenges in multiple fields such as text mining and graph mining and analytics. Text mining techniques must be researched to identify mentions of scientists and organisations (entity linking), but also to detect most informative scientific statements in both scientific publications and text produced by the lay person. Furthermore, tracking of these claims over vast amounts of web content is a challenge that needs to be addressed. The second class of challenges we identify are in the field of network analysis. While the text analysis allows us to link texts to scientists, and to create networks of scientific claims, research must be done in order to understand how these complex networks can be most effectively and efficiently used for the purpose of altmetrics. The current context of social media, mainstream news, the ease of access to government documents, as well as to massive organised collections of scientific publications brings the opportunity to study the complex links and connections between people (journalists, politicians, scientists) at the same time as the links between the content they produce (hyperlinks, paraphrases). This is a great opportunity for understanding scientific impact and scientific finding diffusion, but it comes with great research challenges and questions, some of which we present in the following section.