The challenge
The deadliest window in a disaster is the first 72 hours — before aid arrives, while families drink whatever water they can find. In a country struck by more than twenty typhoons a year, that window opens again and again.
Reactive relief is the most expensive and least effective kind. By the time supplies are mobilized, outbreaks have often already begun. The information to act earlier exists — it just isn’t being turned into decisions fast enough.
Our solution
We’re building an anticipatory-action app: a machine-learning model that takes storm forecasts and community data and predicts where safe water will fail and how much will be needed — before landfall. Humanitarian research is clear that acting before a disaster multiplies the impact of every dollar.
The prototype is live and evolving. Our data scientists work with the Global Centre on Disaster Risk and Poverty, with a path to bring in universities (Ateneo, UP Cebu) and responders like Manila Water and the Philippine Disaster Resilience Foundation.
Who does what
We are force multipliers, not occupiers — the local partner leads and we supply the science, the design, and the funding.
- Build and train the prediction model
- Translate forecasts into water-need estimates
- Design the tool with responders, not just for them
- Keep it open and free for the agencies that act
- Provide disaster and field expertise (GCDRP)
- Ground-truth predictions against real events
- Connect the tool to on-the-ground response networks
The team behind it
Volunteer scientists and engineers donating their expertise to this project — click anyone to see their work.
