Would disaster relief be delivered more efficiently if computers could predict which areas would be worst affected? The Netherlands Red Cross’ new initiative, 510 Global, believes that artificial intelligence could better allocate aid and resources in times of crisis by relying on open data on wind speeds, rainfall and areas previously affected by natural disasters to create a Priority Index that will allow it to dispatch aid where it’s most needed, in the shortest time.
According to a report released by the United Nations Officer for Disaster Risk Reduction, the Philippines was the fourth most frequent country to be hit by natural disasters over the last 20 years. The report, titled The Human Cost of Weather Related Disasters, indicated that a total of 274 disasters hit the country between 1995 and 2015. The UNISDR’s 2015 Global Assessment Report on Disaster Risk Reduction also highlighted that the Philippines is the world’s fifth most vulnerable country in terms of disaster risk implications for development capacity.
By combining data sets, the Red Cross team were quickly able to analyze which areas of the Philippines were most desperate for immediate aid when Typhoon Haima hit the islands last October. In just 24 hours, the 510 Global initiative was able to produce a Priority Index, ranking areas based on need. Data on buildings and infrastructure damaged by the typhoon was later provided by the Department of Social Welfare and Development (DSWD) and the National Disaster Risk Reduction and Management Council (NDRRMC) and proved the Priority Index’s predictions on the worst hit areas correct.
“Our approach is not to develop sophisticated hydrologic, seismic, or windspeed models, but to use machine learning methods to find the best predictors in existing base line data to predict typhoon impact,” writes the team about the on-going development and deployment of artificial intelligence in the 510 Global project. “From our work so far we can conclude that when data preparedness is done right, and disaster impact data collected structurally after an event, then it is possible to use machine learning techniques to build reliable damage predictions. Although damage predictions by using data are not perfect, they are far more transparent than other prioritization methods, because the underlying data, assumptions and methodologies are shared openly.”
Applied research on 510 Global’s objective is ongoing for typhoons in the Philippines, Earthquakes in Nepal and floods in Malawi.