Thus, the tool enables stakeholders to discuss early on, what the intended and unintended effects of the developments may be. The tool utilizes a purely content-based methodology to rank a set of publicly available documents (news articles, blog posts, abstracts from scientific journals etc.) according to their potential to generate risks/impact. The tool considers the textual data over a period of time for a given topic, uses this information to perform data analytics to rank the relevance or importance of the documents and related keywords based on document centrality. Centrality is a simple indicator for gauging the impact of the document that captures the frequency of the terms, but also how these terms are related to each other.
The tool yields the Radar and a Network map. The former includes a list of risks-related keywords on a radar-like widget, along with the list of sources used for extracting and ranking these keywords. The position of the keywords (shown in yellow dots) on the radar indicates its importance where the closer to the centre the term is, the more “relevant” a keyword is in documents related to the COVID-19 coronavirus analysis. The Network map includes a short description of the topic (shown in green circle) connected to the most common keywords (shown in blue circles) and the articles (in colors of red, orange and yellow). The color and size of the collected articles are designated based on their relevance; the larger circle has the highest applicability to the topic.
By looking at different sources (e.g. “science, “media” and/or “public”) the system identifies the topics of interests in the information parsed and assigns them the respective risk scores according to the criteria set (e.g. sentiment, credibility of source, context, etc.). The early warning is identified on the basis of the relation of the terms in search, not on simple “word bagging”. The tool allows user to drill-down to the initial source of information and obtain the background information (e.g. the full-text report).