Inspiration
We were deeply moved by the stories of patients—especially those with Thalassemia, who depend on timely blood transfusions. Despite many well-meaning donation drives, the lack of predictive tools in blood banks often leads to critical shortages. Inspired by how AI can be used in public health, we envisioned a system that doesn’t just track blood, but predicts demand, informs staff, and activates donors before a crisis occurs.
What it does
BloodLineAI is a predictive blood inventory management platform that: Uses machine learning to forecast demand for various blood groups Integrates external data like weather, festivals, and traffic incidents to improve prediction accuracy Sends alerts to hospital staff and donor coordinators when shortages are projected Automatically engages donors via SMS or email Visualizes current inventory and upcoming needs in an intuitive dashboard
How we built it
We will built BloodLineAI using a modular and scalable tech stack: Backend: Flask (Python) for handling APIs, with PostgreSQL for data storage ML Models: Prophet for time-series forecasting TensorFlow for deep learning extensions Frontend: React.js with simple components to render predictive dashboards Deployment: Dockerized backend running on AWS EC2 Data Integration: Simulated data streams from hospital APIs, weather APIs, and donor lists All demand prediction math was handled with models
Challenges we ran into
Real-world data is messy — missing, sparse, or inconsistent Simulating realistic patterns of seasonal demand and donation rates Ensuring data privacy and security while simulating hospital APIs Managing time constraints while integrating multiple modules (backend + ML + frontend) Making the user interface intuitive for medical staff, not just developers
Accomplishments that we're proud of
What we learned
AI can be deeply impactful in healthcare when paired with human-centered design
What's next for BloodLineAI
Partner with blood banks and hospitals for real data integration Improve predictive accuracy in low-data regions using transfer learning Build mobile-friendly, multilingual interfaces for rural access Enable geofenced alerts to nearby donors using GPS APIs Collaborate with NGOs and government for large-scale adoption
Built With
- docker-aws-ec2-(demo-deployment)
- figma
- flask
- postgresql
- prophet-react.js
- python
- tensorflow
- twilio
