Disaster management using advanced analytics

Imagine this,

It’s November. A super-typhoon is bearing down on the Philippines. Local authorities have 48 hours to decide which of 12 million coastal residents to evacuate — and which roads, shelters, and supply routes are still viable. The tools they have? Spreadsheets. Radio. And instinct.

Across South East Asia, natural disasters kill tens of thousands of people each year — not because the data doesn’t exist, but because decision-makers can’t access, interpret, or act on it fast enough.

This project set out to change that — designing a disaster analytics platform that transforms satellite imagery, sensor networks, and historical incident data into clear, actionable intelligence for emergency coordinators in the critical hours before and after a disaster strikes.

How does this work?

The platform aggregates real-time data streams — weather forecasting models, satellite imagery, road and shelter capacity data, and historical disaster records — into a single coordinated view. Machine learning models developed with Google and Microsoft identify likely impact zones and prioritise evacuation corridors.

The UX challenge was to translate this complexity into something that a local government officer — under extreme stress, in a potentially degraded network environment — could act on within seconds. Every design decision was tested against field scenarios with real emergency coordinators across the Philippines, Indonesia, and Vietnam.

Life-Saving Data Visualisation

Information architecture was designed around the three critical decision windows: 72-hour pre-landfall preparation, immediate response in the first 24 hours, and 7-day recovery coordination.

Coming Soon

• Research Process
• Testing
• Implementation and Impact

*Full case study details can be shared during an interview or portfolio review.