Combining Twitter Data, Citizen Science, and Artificial Intelligence to Improve Early Flood-Warning Systems

60 views Leave a comment

Scientists have shown that sea levels have been rising during a rate of about 3.4 mm per year, melancholy to make a continue anomalies of now a normal of a future.

With flooding events augmenting in magnitude and extremity, early-detection and warning systems are apropos some-more and some-more applicable to coastal populations around a world.

Unfortunately, stream methods of detection, such as remote satellite sensing, internal sensor networks, declare statements, and word reports are too delayed and dangerous to be of most use in disaster-prevention.

To this end, Dr Roger Wang and colleagues from a University of Dundee’s School of Science and Engineering in Scotland had total Twitter data, citizen reports and synthetic comprehension algorithms to indication flooding as tighten to real-time as possible.

Researchers used pivotal difference used in Twitter messages, healthy denunciation estimate to accumulate some-more information on astringency and location, and mechanism prophesy to analyse cinema uploaded by users of a crowdsourcing app MyCoast to demeanour for signs of flooding.

“We found these big-data formed inundate monitoring approaches can really element a existent means of information collection and denote good guarantee for improving monitoring and warnings in future,” pronounced Wang in a press release.

Low-lying cities and states are during a larger risk of flooding, necessitating modernized early-warning systems. Image credit: US Army Corps of Engineers around, CC0 Public Domain.

Precision of a information extracted by a algorithm was certified by comparing it to inundate information (which correlated with flood-related tweets) and highway closure reports (found to compare with a information collected from a app).

Wang recognized a need to serve rise a mechanism prophesy techniques used in a investigate – now during roughly 70% correctness – though claimed that Twitter could be used to accumulate coarse-grained information on large-scale events, while information supposing by “citizen scientists” is potentially useful on a micro level.

“Taken together, these collection can be used to guard a H2O invasion of civic flooding over a city. This can be afterwards used to urge forecasting models and early warning systems to assistance residents and authorities ready for an arriving flood.”

The paper was published in a latest book of a biography Computers Geosciences.


Comment this news or article