AI-Powered Predictions for Diabetes & Heart Disease

HealthPulse uses machine learning models trained on clinical datasets to estimate your risk from a few health measurements. It is free, instant and private.

82%Diabetes model accuracy
79%Heart disease model accuracy
A person reviewing a health report on the HealthPulse dashboard
Heart checkResult in seconds
Diabetes checkFree for every user

Built with

  • Google Colab
  • Flask
  • Python
  • MongoDB
  • Chart.js

Two health checks, one account

Pick the assessment you need. Each one has its own model, trained and tested separately.

Diabetes prediction

Enter values such as glucose, blood pressure, BMI and age to see whether your profile points to a higher risk of diabetes.

  • Trained with Scikit-Learn on a labelled patient dataset
  • Clear result with your input values saved to your history
82% accuracy on the test set
Check diabetes risk

Heart disease prediction

Share details like age, cholesterol, resting blood pressure and chest pain type to estimate your likelihood of heart disease.

  • Separate model tuned for cardiac risk factors
  • Charts on your dashboard show how results change over time
79% accuracy on the test set
Check heart risk

Three steps to your result

No lab visit and no paperwork. Keep your latest health report nearby and you are ready.

1

Create a free account

Sign up with your email in a few seconds. Your profile keeps every assessment in one place so you can compare results later.

2

Enter your health details

Fill in a short form with measurements such as glucose level, blood pressure, cholesterol and BMI. Each field explains what it means.

3

Get your prediction instantly

The model returns a result in real time, shown on your dashboard with charts. Save it, track it, and share it with your doctor.

Tested on data the models had never seen

Both models were trained with Python and Scikit-Learn, then scored on a held-out test dataset. These are the accuracies we measured.

Python + Scikit-LearnModel training and evaluation
Flask + MongoDBBackend and user data storage
Diabetes prediction82%
Heart disease prediction79%

Accuracy shows how often a model's prediction matched the test labels. It is not a guarantee for any single person.

Everything you need to understand your numbers

HealthPulse is more than a one-time calculator. It gives you tools to follow your health over time.

Real-time predictions

Submit your details and see the result immediately, with no waiting and no email back and forth.

Interactive dashboard

Charts built with Chart.js show your past assessments so you can spot changes at a glance.

Personal profile

Keep your profile and history together. Update your details whenever your measurements change.

Saved in MongoDB

Your assessments are stored securely in your account so you can come back to them any time.

Works on any screen

Use HealthPulse on your phone at the clinic or on a laptop at home. The layout adapts to both.

Free to use

There are no fees and no premium tier. Register once and run as many assessments as you need.

Know your risk today, talk to your doctor tomorrow

Early awareness helps you act sooner. Create your free account and run your first assessment now.

Healthcare background used on the HealthPulse project

About the HealthPulse project

HealthPulse is a machine learning project that predicts diabetes and heart disease. It combines trained models, a Flask backend and a live dashboard in one web app.

  • Built ML models for diabetes and heart disease prediction using Python and Scikit-Learn.
  • Reached 82% accuracy for diabetes and 79% for heart disease on the test dataset.
  • Designed a Flask-based backend with MongoDB integration.
  • Deployed an interactive dashboard for real-time predictions.

Questions people ask before they start

Is HealthPulse a medical diagnosis?

No. It gives a risk estimate from a machine learning model. Only a doctor can diagnose a condition, so please discuss your result with one.

How accurate are the predictions?

On our test dataset the diabetes model reached 82% accuracy and the heart disease model reached 79%. Real-world results can differ for individuals.

What information do I need to enter?

Common clinical measurements such as age, blood pressure, glucose, cholesterol and BMI. Each form field explains what to enter.

Does it cost anything?

No. HealthPulse is free. You only need an account to save and view your results.