SangBandhan

SangBandhan – Binding Lives with Every Drop.

Comment

About the Project – SangBandhan

Inspiration

Every year, thousands of children in India are born with Thalassemia, a genetic blood disorder that demands lifelong blood transfusions and specialized care. One of our team members had a close friend who struggled with timely access to compatible blood. This personal connection led us to research deeper, where we discovered:

  • Blood donors drop off frequently
  • Delays in matching patients with donors can be fatal
  • Patients and families lack access to proper disease education
  • NGO efforts (like Blood Warriors) are commendable but lack technological scale

Thus, we created SangBandhan – a project born from compassion, shaped by technology, and aimed at binding lives through timely blood, care, and education.

What We Learned

Throughout this journey, we learned to:

  • Implement real-world AI for donor behavior prediction and care personalization.
  • Work with Firebase Cloud Functions and Firestore for real-time, scalable backends.
  • Integrate Google Maps and geolocation APIs into healthcare-focused UX flows.
  • Understand healthcare challenges, especially around data privacy, system trust, and ethical scraping.
  • Collaborate under pressure and design for social impact rather than just functionality.

How We Built SangBandhan

Our tech stack was chosen for speed, scale, and accessibility:

  • Frontend: FireBase Studio
  • Backend: Node.js and Firebase Functions (for API orchestration and ML handling)
  • AI/ML: Python with scikit-learn for donor prediction + LangChain + Gemini for our multilingual Thalassemia awareness bot
  • Database: Firebase Firestore for real-time updates and user data
  • Maps & Matching: Google Maps API to locate donors, blood banks, and match regions
  • External Sync: Simulated scraping + planned integration with e-RaktKosh and Blood Bridge

[ \text{Probability}_{\text{next donation}} = \sigma(\mathbf{w}^T \cdot \mathbf{x} + b) ] Where:

  • ( \mathbf{x} ) = feature vector of past donation activity
  • ( \sigma ) = sigmoid activation (for classification)
  • ( \mathbf{w} ), ( b ) = model weights, bias

Challenges We Faced

  • No official API access to e-RaktKosh – Solved by simulating/scraping data carefully for demo purposes
  • Matching AI logic to real-life donor behavior – Created a hybrid model using donation patterns + constraints (e.g., 90-day wait)
  • Ensuring accessibility in low-connectivity regions – Implemented offline support using Firestore’s local persistence
  • Training a medical chatbot – Had to fine-tune responses using verified sources to maintain medical accuracy
  • Role-based complexity – Managing roles for donors, patients, doctors, and admins required strict auth + conditional UI rendering

What we learned

  • How to combine AI + empathy to build healthcare tools that truly help people
  • Leveraging Firebase for fast, secure, and scalable app development
  • Realizing the value of domain research — we studied WHO, Red Cross, and Blood Warriors' resources to guide our solution
  • Handling health-related data ethically and responsibly
  • Building for multi-role systems and community use cases — not just individual users

What's next for SangBandhan

  • Official integration with e-RaktKosh and Blood Bridge (post-hackathon outreach)
  • Voice-based assistant for low-literacy patients in rural areas
  • Multilingual interface to break language barriers across states
  • Genetic counseling & awareness modules to reduce new Thalassemia births
  • NGO + Hospital onboarding portal to grow the ecosystem
  • CSR/Donor Portal for organizations to support drives and patients ## Final Thoughts

SangBandhan is not just a hackathon project — it’s a step toward a more compassionate, connected, and intelligent health ecosystem. We believe that with the right technology and purpose, we can save lives and build stronger communities, one drop at a time.

“In every drop of blood, there’s a story of hope. SangBandhan exists to make sure it’s heard.”

Built With

  • and-web-`react.js`-?-admin-dashboard-with-analytics-###-??-backend-&-serverless-`node.js`-?-core-logic-`firebase-cloud-functions`-?-api-handling-&-ml-orchestration-###-?-authentication-&-hosting-`firebase-auth`-?-user-login-(patients
  • auth
  • doctors
  • donors
  • e-raktkoshapi
  • firebase
  • firebasestudio
  • flutter
  • google-maps
  • google-places
  • ios
  • langchain
  • react
  • scikit-learn
Share this project:

Updates

Submission history