Diabetes Prediction & Analysis App

A Streamlit-based machine learning application for diabetes prediction and exploratory data analysis. The project uses Logistic Regression and K-Nearest Neighbors (KNN) models and provides an interactive interface for exploring the dataset and generating prediction results.

---

Project Overview

This project combines exploratory data analysis, data preprocessing, machine learning, and interactive visualization into a single Streamlit application.

The application allows users to explore the diabetes dataset, analyze relationships between different features, and use trained classification models to generate diabetes predictions based on entered health information.

---

Features

• Dataset overview
• Dataset shape and statistical summary
• Missing-value analysis
• Exploratory Data Analysis (EDA)
• Count plots
• Box plots
• KDE plots
• Correlation heatmap
• Diabetes prediction using Logistic Regression
• Diabetes prediction using K-Nearest Neighbors (KNN)
• User-friendly prediction interface
• Prediction probability and result display

---

How It Works

The application follows a machine learning workflow:

Diabetes Dataset
      ↓
Data Exploration
      ↓
Data Preprocessing
      ↓
Exploratory Data Analysis
      ↓
Feature Preparation
      ↓
Machine Learning Models
      ↓
Prediction
      ↓
Prediction Probability & Result

The application provides two classification approaches:

• Logistic Regression
• K-Nearest Neighbors (KNN)

---

Dataset

The project uses diabetes_prediction_dataset.csv.

The dataset contains health and demographic information used for analysis and prediction, including:

Feature	Description
gender	Gender of the patient
age	Age of the patient
hypertension	Hypertension status
heart_disease	Heart disease status
smoking_history	Smoking history
bmi	Body Mass Index
HbA1c_level	HbA1c level
blood_glucose_level	Blood glucose level
diabetes	Diabetes outcome

---

Tech Stack

Technology	Purpose
Python	Application and machine learning development
Pandas	Data processing and analysis
NumPy	Numerical operations
Scikit-Learn	Machine learning and model development
Matplotlib	Data visualization
Seaborn	Statistical visualization
Streamlit	Interactive web application

---

Project Structure

diabetes-streamlit-app/
│
├── assets/
│
├── app.py
├── diabetes_prediction_dataset.csv
├── requirements.txt
├── .gitignore
└── README.md

The exact repository structure may vary depending on the current project files.

---

Getting Started

Clone the Repository

git clone https://github.com/devsparkcodes/diabetes-streamlit-app.git

Navigate to the Project

cd diabetes-streamlit-app

Create a Virtual Environment

python -m venv venv

Activate the Virtual Environment

Windows:

venv\Scripts�ctivate

macOS / Linux:

source venv/bin/activate

Install Dependencies

pip install -r requirements.txt

Run the Application

streamlit run app.py

The application will open through the local Streamlit server.

---

Learning Outcomes

Through this project, I practiced:

• Data exploration and analysis
• Data preprocessing
• Exploratory Data Analysis
• Feature preparation
• Classification models
• Logistic Regression
• K-Nearest Neighbors
• Model prediction
• Data visualization
• Building interactive Streamlit applications

---

Future Improvements

• Compare additional machine learning models
• Add detailed model performance comparison
• Improve prediction visualizations
• Add model explainability
• Improve the overall user interface
• Add more comprehensive evaluation metrics

---

Disclaimer

This project is intended for educational and demonstration purposes only. Its predictions should not be considered a medical diagnosis or a substitute for professional medical advice.

---

Author

Muhammad Umar

Building practical applications at the intersection of software engineering and AI.

• GitHub: https://github.com/devsparkcodes
• LinkedIn: https://linkedin.com/in/devsparkcodes