Officers responsible of crisis management and security assessment need accurate timely information on the security situation on the field.
Textual news data coming from various sources (news agencies, government reports etc.) need precise categorisation, geo-location, extraction of meta-data and a fast and easy visualisation to be effective.
We aim for the creation of a pipeline for information extraction, categorisation and the final visualisation on an interactive dashboard, to be used for the security of Non-Governmental Organisation Personnel and Mission members.
Similarly, we work onto support to WHO’s Epidemics Intelligence from Open Sources to apply the latest Text Mining techniques to detect and tack future epidemics.
- Selection of trusted input sources and definition of a standard metadata structure.
- Implementation of an analysis pipeline that takes the news article as input and outputs the set of relevant metadata.
- Studying and implementing the best techniques for noise reduction, data validation, data classification and data clustering.
- Implementation of an interactive graphical interface for an easy access to the relevant information.
- Applied research on geocoding, event extraction, (sub)-topic classifiers.