Inspiration

Singapore experiences intense, short-duration rainfall that can quickly lead to flash floods, disrupting roads, public transport, businesses, and daily life. What if we could predict where flooding is likely to happen before it actually does? Existing flood alerts are useful for responding to flooding, but they are largely reactive. RainCheck takes a proactive approach by using rainfall patterns and machine learning to identify areas at risk before flooding occurs.

What it does

RainCheck combines historical rainfall, real-time rainfall, and historical flood alert data to predict flood risk across Singapore. The system classifies geographical zones into Low, Medium, or High risk, alongside confidence scores, and presents the results through a live dashboard. The goal is to give residents, transport operators, and authorities more time to prepare, reroute, evacuate, or mobilise resources before flooding occurs.

How we built it

Our architecture sources historical and real-time rainfall and flood alert data from Singapore's data.gov.sg APIs. The data is processed and analysed to identify rainfall patterns associated with flooding. We then use machine learning to learn from historical trends and predict future flood risk based on incoming rainfall data. These predictions are presented through a live dashboard, allowing users to see current rainfall conditions and the predicted risk across Singapore.