New tool to predict floods in real-time now available online
Floods are usually detected and monitored by using hydrological models or satellite imagery. A tool developed by researchers at IVM-VU in collaboration with FloodTags, Deltares and the Radboud University now provides an additional indication for the development of floods using data from Twitter. The tool, called the Global Flood Monitor, is aimed at helping first-responders act quicker in case of emergency, and is now available online.
Global Flood Monitor
The Global Flood Monitor uses Twitter data to determine the location of floods world-wide and generates a world-map visualizing relevant tweets in real-time. This information can be very useful as an early warning system for international help organisations (e.g. the Red Cross) and supports the ability of such organisations to respond quicker to calamities as they occur. The platform also makes it possible to access historical data, providing an archive of flood-related Tweets dating back three years. This information on historic events is very valuable, for instance to validate alternative models used in flood detection.
Linking a Tweet to its location
Determining the locations mentioned in tweets is not as straight forward as it sounds. To be able to link individual tweets with confidence to a particular location, required the development of a special algorithm. In an earlier article on this site, you can find more information on the work done by Jens de Bruijn in collaboration with his colleagues at FloodTags and the Philippine Red Cross.
After the mentioned locations are extracted from the tweets, the algorithm keeps track of the number of mentions of all places within a region. When a region is mentioned more often than normally, it is classified as a flood.
The Global Flood Monitor is met with great interest from global institutions such as the Red Cross and World Bank, and by government organisations. Also, the new algorithm holds great promise to also be of use in areas other than detecting floods; for example, the prediction of droughts or the occurrence of harmful algal blooms.
Twitter, Instagram and Facebook
Additionally, the developers want to include data from other types of social media, such as Facebook, Instagram and blog posts. This may be particularly interesting because of the additional image data that is available through these media.
The work by Jens de Bruijn and colleagues was recently published in the Journal of Geovisualization and Spatial Analysis.
Bruijn, J. De, Moel, H. De, Jongman, B., Wagemaker, J., & Aerts, J. C. J. H. (2018). TAGGS : Grouping Tweets to Improve Global Geotagging for Disaster Response. Journal of Geovisualization and Spatial Analysis.
On the launch of the tool, Dutch BNR News Radio interviewed Jeroen Aerts (Professor Water and Climate Risk, IVM-VU).
Jens de Bruijn
Jens de Bruijn works as a PhD Candidate at the Institute for Environmental Studies (IVM), Vrije Universiteit Amsterdam. His research focusses on the application of various social media in flood risk modelling, (near-)real time flood modelling and flood predictions. The research is conducted under the supervision of Dr Brenden Jongman, Dr Hans de Moel and Prof. Jeroen Aerts. Jens is part of Amsterdam Water Science.