3 questions to...
Marc Rigolot Director of the MAIF Foundation for Research
What brought the MAIF Foundation and BRGM together?
M. R. — A longstanding and historic partnership has been forged between the MAIF Foundation and BRGM. As a public interest foundation, we work for the general good of society. Moreover, given that our primary activity is insurance, we are particularly concerned by risk.
We fund applied research to further knowledge and prevention of risks, meaning that all our projects have a direct impact on society. Our objectives are to reduce risks and take measures to limit accidents, in everyday life, in transport or involving natural hazards.
BRGM’s teams have recognised expertise in this field. We were therefore keen to support the SURICATE-Nat public platform project, which continuously collects and analyses messages published on the social network Twitter to speed up the flow of information from the field when natural disasters occur.
How did this cooperation work in practice?
M. R. — In scientific terms, it involved applying Artificial Intelligence and Big Data to detect an event or measure its magnitude. With the help of the Technological University of Troyes (UTT), BRGM developed a tool for monitoring Twitter content on risks such as earthquakes and floods. Relevant tweets are targeted on a semantic basis (using risk-related keywords) and AI algorithms analyse each message individually: any tweets related to a natural disaster phenomenon are geolocated and classified.
How would you sum up its results?
M. R. — Here we are at the heart of citizen science, using observations from “human sensors” to analyse and deduce what is happening in the field. This has been a really innovative project in the field of natural risks. It ran from 2016 to late 2019, with total funding of €418K, of which €150K was provided by the MAIF Foundation. For such an innovative, applied topic with such an immediate impact, we really appreciated BRGM’s support and responsiveness, as well as its risk expertise and the cohesion of its teams. BRGM was also able to adapt the project according to the obstacles encountered.
The project has now come to an end for us: it covered earthquakes and floods, effectively handled the real-time observation of tweets, and everything is available online. BRGM is taking it further now, in order to improve the reliability and quality of the diagnostics and make it available for other risks. This is just the start of the creation of communities to inform the authorities and prevent accidents.