Inspiration
Millions of people are personally affected by climate disasters every year. And in disaster response, every second counts. When healthcare systems are forced to react instead of anticipate, lives are lost in the delay. Our project focuses on bridging the gap between climate data and health preparedness by transforming historical disaster data into actionable insights that aren’t just innovative but lifesaving.
How we built it
We used historical natural disaster data that described disaster types by region, location, frequency and corresponding health impacts. After importing the dataset, we cleaned and standardized it to omit excess info and check for missing values. We then trained an AI model to create a predictive seasonal forecast. Our resulting solution was an interactive analytical interface that allows users to explore disaster patterns spatially, temporally, and seasonally.
Next, we built a global geo heatmap that visualizes the average number of natural disasters by location, making it easy to identify high-risk regions at a glance.
We then added a historical seasonal trendline, showing how disaster frequency changes throughout the year. This helps highlight seasonal risk patterns that are especially relevant for health preparedness.
Building on this historical analysis, we developed a predictive seasonal forecasting model. Using this model, we generated a forecasted trendline and directly compared it against historical trends to show how disaster patterns may evolve in the future. We achieved an accuracy of 87 percent with our models.
All of these visualizations are integrated into a single interactive interface, allowing users to filter, compare, and explore insights dynamically rather than relying on static reports.
What's next for Climate Disaster and Global Health Impact
Our team sees this application being used in the real world by public health agencies, humanitarian organisations and hospitals worldwide to anticipate periods of heightened risk and effectively allocate resources well in advance of disasters. In the future, with larger datasets and more accurate modeling, this platform has the potential to change how lives are protected during disasters — because behind every data point is a real person, and every improvement lives saved.
Log in or sign up for Devpost to join the conversation.