A multiple disease prediction web application built using machine learning provides a fast, accessible and interactive way for early health screening.
This project allows users to input medical parameters and instantly receive predictions for three major conditions
– Diabetes
– Heart disease
– Parkinson’s disease
In a world where early awareness can significantly improve health outcomes, having a lightweight digital tool that offers instant health insights is extremely valuable. Machine learning enables this by analyzing patterns from medical datasets and delivering quick predictions through a simple interface.
Using Python and Streamlit the system delivers a smooth browser based application capable of real time predictions using three separate trained machine learning models.
The goal of this application is not to replace medical diagnosis, but to support early awareness and encourage users to take proactive steps toward their health.
Project Overview
Machine learning powered multiple disease prediction system
Integrated into a Streamlit web app with sidebar navigation for easy module switching
Uses three pre trained models loaded from pickle files
Accepts disease specific input values and returns an instant prediction for each condition
Models included
Diabetes prediction model
Heart disease prediction model
Parkinson’s disease prediction model
Libraries Used
NumPy for numerical array handling
Pickle for loading the pre trained ML models
Streamlit for building the web interface
Streamlit Option Menu for sidebar navigation
Scikit Learn models used offline before exporting to pickle files
Application Details
The project loads three separate models stored as diabetes_model.sav, heartdisease_model.sav, and parkinson_model.sav
Each disease prediction section has its own user input fields displayed in organized columns, allowing structured and readable data entry.
1 Diabetes Prediction Inputs
Number of pregnancies
Glucose level
Blood pressure
Skin thickness
Insulin level
BMI
Diabetes pedigree function
Age
2 Heart Disease Prediction Inputs
Age
Sex
Chest pain type
Resting blood pressure
Cholesterol
Fasting blood sugar
Resting ECG results
Maximum heart rate
Exercise induced angina
Oldpeak
Slope
Major vessels count (ca)
Thal value
3 Parkinson’s Disease Inputs
A set of 22 vocal frequency and biomedical parameters including MDVP measures, jitter values, shimmer values, NHR, HNR, RPDE, DFA, spread measures, D2 and PPE
Streamlit : Diabetes Prediction
User values collected through Streamlit text inputs
Inputs converted into numeric form when required
Values structured into a list and then into an array
Array reshaped properly so the ML model understands it is a single prediction case
Model returns a binary output where
Zero means the person does not have the disease
One means the person has the disease
Model Interaction
Each model is loaded using pickle load at app start
When user clicks the prediction button the corresponding function
Collects inputs
Converts and reshapes values
Feeds them into the correct model
Displays a clean readable message to the user
Example logic used
If prediction equals one return person has the disease
Otherwise return person does not have the disease
Streamlit Interface
A sidebar menu allows switching between Diabetes Heart Disease and Parkinson’s Prediction
Each disease module has its own section title and organized input layout
Prediction results appear using st.success for clean and visible output
Error messages are shown when non numeric values are entered in disease modules that require numeric input
The interface is simple minimal and effective for fast testing.
Prediction Flow
1 User selects disease type from sidebar
2 User enters medical values
3 Inputs are validated and converted
4 ML model processes the values
5 Result is shown immediately on screen
Deployment Possibilities
Deploy easily on Streamlit Cloud for public use
Package into a Flask or FastAPI project for full stack deployment
Use inside community health tools for awareness sessions
Add authentication or admin dashboards for clinics
Key Takeaways
A complete multi disease prediction web application using machine learning
Demonstrates how multiple trained models can run inside a single interactive interface
Shows real world use of ML for health screening and awareness
Easy to extend with more diseases, better UI components and improved analytics
Future Enhancements
Add automatic validation for all numeric inputs
Replace text fields with sliders or dropdowns for better accuracy
Add visual graphs showing input distributions or health indicators
Include explanations of predictions using SHAP or similar tools
Make interface mobile friendly and responsive
This project highlights how machine learning and simple web frameworks can come together to create powerful tools that improve awareness and accessibility in healthcare. By allowing users to check for diabetes, heart disease or Parkinson’s disease within seconds, this application demonstrates the impact of combining technology with health driven use cases. It also shows how multi model systems can be packaged into a single unified interface, making it an excellent learning project for anyone exploring data science, ML deployment or health informatics.