25/12/2015

Artificial Intelligence for Disaster Response (AIDR)

 

A tool that harnesses artificial intelligence to mine tweets in humanitarian crises.

Increasingly, social media is at the forefront of disaster reporting. Eye-witnesses post photos, cries for help, and situational information in real-time on networks like Twitter. Yet, given the vast number of tweets published, for journalists to truly harness Twitter as an information source, they would need to employ some sort of social media mining tool in order to identify the pertinent information.

In comes Artificial Intelligence for Disaster Response (AIDR) - a tool developed by the Qatar Computing Research Institute (QCRI), that utilizes machine learning to automatically identify tweets related to a particular crisis.

AIDR is made up of two key components: the collector and the tagger. Firstly, the collector mines tweets related to a disaster based on your selection of keywords, geolocation and hashtags. The tagger then learns to classify tweets by topic - say, road blockage or water contamination - from similar tags crowdsourced through MicroMappers.org.

aidr.png

Some examples of disasters that the AIDR platform has been utilized to report on include the MH370 disappearance and the 2014 Chile Earthquake.

Check out the following video for a short tutorial on how to use the AIDR platform:

Visit the AIDR website here or its GitHub page here.

Photo: Spencer E Holtaway

Comments