Flooding has the potential to drastically damage critical infrastructure and impact human mobility, especially in underserved communities. So far, in 2017, flooding has been responsible for half of the deaths and half of the economic damages from climate-related disasters.

Houston is just the latest example.  A recent project produced by Ph.D. students Amrita Gupta and Caleb Robinson, along with School of Computational Science and Engineering Assistant Professor Bistra Dilkina, aims to alleviate and address this issue by outlining a framework needed to develop and maintain resilient road networks, while supporting multiple sustainable goals. The team also enlisted the help of Lingzi Hong, Ricardo Macias, Brendon Machado and Keyan Halperin, participants in the Data Science for Social Good (DSSG-ATL) summer program held at Georgia Tech.

The project – Predicting and Alleviating Road Flooding for Climate Mitigation – won the Thematic Award in Climate Adaptation for the United Nations (UN) Data for Climate Action Challenge, an unprecedented open data innovation competition to leverage big data and analytics for social good. The official announcement was made Sunday, Nov. 12 at the UN Sustainability Innovation Forum. Robinson traveled to Bonn, Germany to receive the award on behalf of the team at the forum, which served as a companion event to the annual UN Conference COP23.

The high-level event featured discussions and examples of data-driven innovations to accelerate climate solutions and the Sustainable Development Goals (SDGs) by speakers from the UN, Western Digital, Twitter, Tableau, World Resources Institute, and more.

"We are always looking for ways to apply methods from computer science to problems in sustainability in our research group. The UN Data for Climate Action Challenge was an amazing opportunity to do just that, with the added benefit of having access to high-quality data that is often difficult to obtain otherwise,” said Gupta.

The project details how to make a computational framework to determine how best to improve the flood resilience of the road networks in an area under different flooding scenarios. The team from Georgia Tech believes this will help support multiple goals including alleviating poverty in low-income countries by reducing disaster risk and by providing reliable access to healthcare, education, and financial services.

According to Dilkina, “This framework can consider different optimization objectives and budget constraints in order to give policy makers, investors, or relief organizations an idea of the trade-offs associated with these parameters. We demonstrate our framework in Senegal, by using flood hazard data from Fathom.Global and mobility data from Orange S.A.”

The framework of the project includes three separate parts:

• Estimate flooding effects on road networks by quantifying how a road network is affected by flooding (using GIS analysis to couple flood hazard raster data and road network data)
• Compute resulting impact on mobility by finding which roads in the network are most used by people (using network analysis techniques to couple cellphone call record data and road network data)
• Recommend road fortifications that most effectively prevent loss of mobility (using optimization techniques)

“It is a great problem that fit well with the data provided in the challenge and our expertise in data science and optimization. Attending the summit in Bonn was extremely inspirational. Talking with the other winning teams and the UN Global Pulse team was a great opportunity, and I learned a lot about projects that are currently underway in the climate action space, as well as exciting new directions,” says Robinson.

View the recording of the team’s presentation at the UN event here: https://www.youtube.com/watch?v=C_rDd1V72AQ&feature=youtu.be&t=54m30s