Semantic Explorer (Semex.io) is a text mining tool (TM) based on cutting edge technology capable of extracting information from unstructured data and display the results on clear graphical and textual outputs. It has been designed with the objective of providing support to researchers involved in Foresight, System Dynamics Modelling and Policy Evaluation by reducing the cognitive load related to tasks that are essential to policy processes. Thus, specific solutions have been developed for this purpose.
The portal is the gateway to a large library of more than 2700 documents related to European rural areas. Sources have been gathered by PoliRural Pilots organizations ensuring a high level of expertise in rural related topics and wide geographical representation (12 Pilots in different parts of Europe). The scientific articles, technical reports and policy related documents collected relate to the needs of PoliRural areas as well as to the local, regional, national and European policies.
In semex.io users can access and process information following different approaches that are summarized on the top menu of the frontend application. It is possible to consult the Library´s contents by browsing Sources and selecting those that one is interested in. Users can work on a single source or on a selection of articles that can be gathered in the Curated Reading List. The tool analyses the text and provides various results such as text summary, topic extraction, most relevant Named Entities, sentiment analysis, geographical parsing etc…. Very soon there will be the possibility of visualizing, out of a large corpus of text such one or more articles, the most recurrent topics or key words through custom scattered plots based on word2vec technique.
It is also possible to access sources through a fuzzy search which includes the possibility of Boolean logic search as for instance ´depopulation´ AND ´rural areas´. Through the Topic option users can explore the GEMET thesaurus and discover relations between topics associated with the rural issues. Topics are automatically connected to Library’s sources and are displayed on the dedicated panel along the text. For data analysts who wish to dig deeper into the data a specific dashboard has been set in Kibana and is accessible for registered users.
All these functionalities have been developed following intense communication with researchers involved in the PoliRural project. The aim is to provide support to researchers and policy-makers in tasks related to policy processes by providing automatized analysis of large corpus of text. The next few months will be dedicated to test the tool and collect feedback with the objective of further tailoring solutions, replicable also for users outside of the PoliRural consortium.