CKD Prediction App

Chronic Kidney Disease Prediction System
Group M - Scientific Programming Final Project (MHEDAS 2025-26)

---

Overview

A machine learning-powered web application for predicting Chronic Kidney Disease (CKD) based on clinical biomarkers. The system uses a K-Nearest Neighbors (KNN) classifier trained on patient data to provide real-time predictions through an intuitive web interface.

Key Features

• ML-Powered Predictions: KNN model with 93.75% accuracy
• Web Interface: User-friendly form for inputting clinical values
• Docker Ready: Containerized for easy deployment
• Model Metrics: View detailed performance metrics

---

Quick Start

Option 1: Using Docker (Recommended)

# Pull and run the container
docker run -d --name ckd-app -p 8080:8000 inigoarriazu/ckd-prediction-app:latest

Then open your browser and go to: http://localhost:8080

Option 2: Run Locally

# Clone the repository
git clone https://github.com/andreaperval-gap/SP_FinalProject.git
cd SP_FinalProject

# Install dependencies
pip install -r requirements.txt

# Run the application
cd app
uvicorn mlapi:app --host 0.0.0.0 --port 8000

Then open your browser and go to: http://localhost:8000

---

Docker Commands

Basic Commands

# Start the container
docker run -d --name ckd-app -p 8080:8000 inigoarriazu/ckd-prediction-app:latest

# Stop the container
docker stop ckd-app

# Remove the container
docker rm ckd-app

# View container logs
docker logs ckd-app

Troubleshooting (Port Already in Use)

# Stop and remove existing container, then restart on a different port
docker stop ckd-app
docker rm ckd-app
docker run -d --name ckd-app -p 9000:8000 inigoarriazu/ckd-prediction-app:latest

Then access at: http://localhost:9000

---

Project Structure

SP_FinalProject/
├── app/
│   ├── mlapi.py              # FastAPI application
│   ├── index.html            # Web interface
│   ├── img/                  # Static images
│   └── models/
│       ├── knn_k5.pkl        # Trained KNN model
│       └── knn_k5.metrics.txt # Model performance metrics
├── docs/
│   └── Project_Report.md     # Full project report  
├── preprocessing.py          # Data preprocessing functions
├── Scientific_Programming.ipynb        # Analysis notebook
├── Scientific_Programming+modelos.ipynb # Models notebook
├── chronic_kindey_disease.csv          # Dataset
├── Dockerfile
├── requirements.txt
└── README.md

---

Clinical Features

The model uses the following 11 clinical biomarkers for prediction:

Feature	Description	Units
Age	Patient age	years
Blood Pressure	Diastolic BP	mmHg
Glucose	Blood glucose (random)	mgs/dl
Urea	Blood urea	mgs/dl
Creatinine	Serum creatinine	mgs/dl
Sodium	Serum sodium	mEq/L
Potassium	Serum potassium	mEq/L
Hemoglobin	Blood hemoglobin	gms
PCV	Packed cell volume	%
WBC	White blood cell count	cells/cumm
RBC Count	Red blood cell count	millions/cmm

---

Model Performance

Metric	Value
Accuracy	93.75%
Precision (CKD)	0.96
Recall (CKD)	0.94
F1-Score (CKD)	0.95

Confusion Matrix:

              Predicted
              NOT CKD   CKD
Actual  NOT CKD   28      2
        CKD        3     47

---

API Endpoints

Endpoint	Method	Description
/	GET	Web interface
/models/	GET	List available models
/metrics/?model_name=<name>	GET	Get model metrics
/predict/	POST	Submit prediction request

---

Team

Group M - Master in Health Data Science (MHEDAS 2025-26)

---

License

This project was developed for educational purposes as part of the Scientific Programming course.

---

Documentation

For more detailed information, see:

• Project Report