HealthPulse

"Beat Diabetes with Personalized Care"

Comment

Inspiration

Diabetes is often called the silent killer in India — millions are living undiagnosed until it’s too late. We were inspired to build HealthPulse because prevention shouldn’t feel complicated or inaccessible. Our idea was simple: combine predictive health insights with personalized Indian context (local diets, lifestyles, and habits) so that people can not only understand their risk but also act on it meaningfully.

What it does

1.HealthPulse is a personalized diabetes risk prediction system designed for India. It: 2.Predicts risk levels (Low / Moderate / High) using health metrics and lifestyle inputs. 3.Generates interactive visualizations (bar & radar charts) for quick insights. 4.Creates a professional PDF health report with charts, recommendations, and doctor referral notes. 5.Delivers region-specific advice — from Karnataka’s Ragi to Maharashtra’s Alphonso Mango — making recommendations culturally relevant.

How we built it

  1. Frontend: Built an intuitive interface using Gradio for instant accessibility.
  2. Machine Learning: A Random Forest Classifier processes medical features, supported by a custom scoring system aligned with WHO & ADA guidelines.
  3. Visualization Layer: Dynamic bar and radar charts highlight where the user stands vs. healthy ranges.
  4. PDF Engine: Automated report generation via ReportLab, formatted for both patients and doctors.
  5. Regional Intelligence: A manually curated India-wide region–food–lifestyle database for personalized recommendations.

Challenges we ran into

1.Lack of real clinical datasets, so we balanced synthetic training with guideline-based scoring. 2.Designing accurate yet user-friendly inputs that non-medical users can understand. 3.Managing India’s regional diversity while keeping recommendations relevant. 4.Building professional PDFs with charts that look clean and readable.

Accomplishments that we're proud of

1.Built a fully functional risk prediction system with real-time charts and PDF reports. 2.Added regional food and lifestyle personalization, making the tool truly unique for Indian users. 3.Designed an end-to-end workflow from input to report, usable in both clinics and community health camps. 4.Created a hackathon-ready solution that blends ML, visualization, and health awareness.

What we learned

1.How to integrate machine learning with domain expertise for healthcare. 2.The importance of UX design in health apps, especially for older or non-technical users. 3.How to build state-based modular data systems for localized recommendations. 4.Practical skills in Gradio, visualization libraries, and PDF generation.

What's next for HealthPulse

  1. Training the model on real-world medical datasets for stronger accuracy.
  2. Expanding coverage to all Indian states with seasonal diet tips.
  3. Adding multilingual support (Hindi, Tamil, Bengali, etc.) to reach more users.
  4. Enabling batch CSV uploads for large-scale health screenings.
  5. Integrating with wearables and IoT devices for real-time risk tracking.

Built With

Share this project:

Updates