ThalaMitra – AI Donor Matching & Engagement (MVP)

• 

  application home page

• 

  notification

• 

  eraktkosh

• 

  donation prediction page

• 

  leader board

• 

  eligibility checker

• 

• 

  flowchart how it works

• 

  logo

• 

  architecture

Inspiration

Thalassemia patients need lifelong blood transfusions, often facing last-minute shortages, rare blood group crises, and donor unavailability. Seeing this gap in real-time donor matching and engagement, we envisioned ThalaMitra — an AI-powered platform blending technology, community outreach, and government integration.

What We Learned

AI for Social Good can directly save lives.

How predictive analytics can forecast donor availability using patterns:

𝑃(Donation)=𝑓(Recency, Frequency, Volume, Time) P(Donation)=f(Recency, Frequency, Volume, Time)

Importance of UX in health tech for low-literacy and rural communities.

Combining offline & online systems for inclusivity.

How We Built It

Frontend: Streamlit prototype (scalable to React/Flutter).

Backend: Python (Scikit-learn, Pandas, NumPy).

Database: CSV demo dataset (Transfusion) → expandable to PostgreSQL/MongoDB.

ML Model: RandomForestClassifier to predict donation likelihood.

APIs: Designed architecture for integration with e-RaktKosh, WhatsApp Business API, and location services.

Engagement: Gamification with badges, donor leaderboard, and certificates.

Challenges Faced

Data Privacy: Ensuring donor and patient data security.

Connectivity: Handling rural/low-network regions via offline “ThalaMitra Boxes.”

Integration: Aligning with government systems like e-RaktKosh.

Behavioural Adoption: Motivating recurring donations through rewards + recognition.

Conclusion: ThalaMitra is more than an app — it’s a movement to connect hearts, powered by AI, community, and compassion.

Built With

• aes
• google-maps-api-cloud-&-hosting:-streamlit-cloud-(prototype)
• https
• javascript-(for-future-react/flutter-expansion)-frameworks-&-libraries:-streamlit
• languages:-python
• numpy-machine-learning:-randomforestclassifier
• pandas
• scalable-to-aws/azure-security:-oauth-2.0
• scikit-learn
• time-series-forecasting-database:-csv-demo-dataset-?-scalable-to-postgresql-/-mongodb-apis:-designed-for-e-raktkosh-api
• whatsapp-business-api