Foto 7

Bellomo, Cresci, Del Vigna, La Polla - A platform for gathering eyewitness reports from social media users in the aftermath of emergencies

Written by

Social sensing is based on the idea that communities or groups of people can provide sets of information similar to that which is obtainable from sensor networks. Emergency management and situation awareness are candidate fields of application for social sensing. 

Nowadays, two different approaches are present in literature: opportunistic and participatory crowdsensing, the former of which intends to detect events and/or gain situation awareness by gathering data ‘on- the-fly’ while the latter ‘hires’ volunteers in order to retrieve valuable information. 

This work aims to create, implement and deploy a platform based on a decision support system for gathering eyewitness reports in the aftermath of an emergency, focusing in particular on earthquakes. These reports can be useful to improve the situation awareness. While doing so, we would like to find out if an approach combining opportunistic and participatory sensing methods is effective. Our system, in fact, detects eyewitnesses exploiting an opportunistic approach and then aims to transform these potential eyewitnesses into volunteers willing to share information. 

The platform retrieves earthquake notifications from an official channel and, immediately after, gathers the messages shared on Twitter for a fixed timeslot. In doing so, we collect messages posted by potential eyewitnesses. Data mining and natural language processing techniques are applied in order to select meaning- ful and comprehensive datasets of tweets. We then concentrate on the filtered dataset in order to try to engage with their authors and obtain, in real time, information and enhance situation awareness. 

Information retrieved by our system can be extremely useful to all the government agencies interested in mitigating the impact of earthquakes, as well as news agencies looking for new information to publish. 

Results collected by our platform are promising and, despite being in its preliminary stages, a combined approach to the search for earthquake eyewitnesses seems possible.