💡 Inspiration
We were inspired by how often outbreaks are detected too late, even when early signals already exist. Most systems rely on confirmed data, but by the time patterns are visible, the spread has already begun.
We wanted to explore a different idea: can we detect risk earlier by combining simple, everyday signals like symptoms, location, and exposure?
🛠️ How We Built It
We built a lightweight system that takes user-reported symptoms and enriches them with contextual information such as environmental conditions and exposure factors.
At the core is a simple, explainable scoring model:
$$ Risk\ Score = w_1 \cdot Symptom\ Severity + w_2 \cdot Exposure\ Risk + w_3 \cdot Environmental\ Risk $$
The frontend was developed using React (Vite + Tailwind), and the system is designed in a modular way so data sources and models can be improved over time.
🚧 Challenges We Faced
One of the biggest challenges was working without access to real, labeled epidemiological data. Instead of forcing a complex machine learning model, we focused on building a system that behaves consistently and logically across realistic scenarios.
Another challenge was balancing simplicity and usefulness—making sure the output is easy to understand while still being meaningful.
📚 What We Learned
We learned how to design AI-driven systems under real-world constraints, especially when data is limited.
We also gained a better understanding of:
- building explainable models
- thinking about public health applications
- designing systems that connect technical outputs with real-world decisions
🧠 Final Thoughts
This project is less about perfect prediction and more about early awareness.
Our goal was to show that even simple, well-structured systems can help people make better decisions when it matters most.
Log in or sign up for Devpost to join the conversation.