Students, DxHub Create App for World Bank to Improve School Safety in Developing Countries
The Cal Poly Digital Transformation Hub (DxHub) powered by Amazon Web Services (AWS) and a team of Cal Poly students designed a mobile app for the World Bank to help improve school safety in developing countries, projecting to reduce costs and time by 50%.
Millions of children’s lives and billions of dollars in infrastructure are at stake, since disasters such as earthquakes and cyclones put more than 1 million school buildings in low- and middle-income countries at risk of collapsing. An estimated 875 million children and teachers could be injured or die in damaged school buildings, according to World Bank figures. In addition, when schools close as a result of disasters, there are indirect losses as students are deprived of learning opportunities, further compounding the impacts.
The cutting-edge technology was developed by the DxHub, using AWS Cloud-based tools, for the World Bank’s Global Program for Safer Schools (under the umbrella of the Global Facility for Disaster Reduction and Recovery). A multidisciplinary team of undergraduate and graduate students, and Professor Franz J. Kurfess worked with DxHub officials to create the mobile phone app design and deep learning computer vision model. The model simplifies how different types of school buildings vulnerable to natural disasters are identified to speed up the risk-informed intervention and investment planning to ultimately make schools safer at scale.
The app design, automated data management solution and deep learning model received international recognition from the Observatory of Public Sector Innovation, a flagship initiative of the Paris-based Organisation for Economic Co-operation and Development. The 60-year-old OECD, with 38 member nations, supports and advises governments on making innovative approaches to transform their public sector.
The solution design is the result of a university-based partnership between the DxHub and the World Bank, one of the world's largest sources of funding and knowledge for developing countries. The World Bank helps developing countries share and apply innovative knowledge and solutions to solve the challenges in every major development area towards ending poverty and shared prosperity, through a wide range of financial products and technical assistance.
The app design and the deep learning computer vision model will help the Global Program for Safer Schools (GPSS) move forward its agenda with scalable and efficient data solutions to facilitate governments’ strategic investment prioritization based on evidence on school building vulnerability.
DxHub and Cal Poly faculty and students were called in to help address the development challenge on school infrastructure baseline data. There’s a lack of baseline data about school building structural characteristics in developing countries and no efficient way to collect, update and manage such a large amount of data at national scale. Specifically, field data collection across large geographic areas and labeling of photographic data is time-consuming and expensive to do across tens of thousands of schools.
The DxHub welcomed the opportunity to work with students to tackle this important and critical challenge.
“Our partnership with the World Bank enables Cal Poly students to innovate and problem solve in ways that have real impact,” said DxHub Director Paul Jurasin. “This app design is just one example of the solutions we are bringing to bear that can make a significant difference in the world, and we greatly appreciate the recognition of our collaboration.”
This month, Kurfess and his computer science class demonstrated the latest iteration of the solution design, which the GPSS team can now use as a template in field testing.
“The partnership with Cal Poly and DxHub helped us explore an AI-based solution to supplement the country-level school infrastructure data gap leveraging a team of talented young minds from multidisciplinary backgrounds,” said Task Team Lead of the GPSS, Fernando Ramirez Cortes. “This approach will not replace the conventional field surveys by qualified personnel but will complement the more detailed assessments over a representative sample. By drawing conclusions at a large scale, the AI solution prevents the need to carry out those conventional assessments over the entire portfolio and is projected to reduce time and expense by 50%.”
This month’s final demonstration of the model culminates two years of research and development by multidisciplinary teams, including a virtual student exchange with the Munich University of Applied Sciences, also part of the AWS Cloud Innovation Centers Program. The teams have examined the use of AI technologies for various aspects of making schools safer, creating an easy-to-use mobile app design to identify vulnerable school buildings most in need of interventions. Using the app, photos of school buildings are uploaded to the cloud where an AI deep learning algorithm determines the key vulnerability characteristics of each structure — building type, height and structural system.
A trained engineer remotely reviews the initial results for accuracy and, in the process, fine-tunes the deep learning model to continuously improve the quality of the results. From there, the engineer’s findings go to decision makers who can quickly prioritize investments for the most vulnerable school buildings.
For students involved, including DxHub student employees, Ashley Choi, a sociology and political science senior from San Mateo, California, and Elisa Horta, a mechanical engineering senior from Santa Clara, California, this project provides a Learn by Doing educational experience, as they graduate, ready to move on to careers in industry.
“This project was a great opportunity to accelerate my learning in a real-world setting,” said Cal Poly alumna Sydney Nguyen (architecture and ethnic studies, ’20) from Oakland, California. “I was especially drawn to the social impact of this project, with learning opportunities to use design and technology to aid marginalized communities at local-global scales.”
In addition to recognition from the OECD’s Observatory of Public Sector Innovation, Kurfess and the team presented at the 128th American Society for Engineering Education Annual Conference and also were recognized as a semifinalist in the SAFE STEPS D-Tech (Disaster Tech) Award for innovative technology solutions that save lives before, during and after natural disaster events.