A machine learning API that predicts diabetes progression scores based on physiological features. Built with FastAPI and scikit-learn, this service provides real-time predictions through a RESTful API interface.

Try the API directly: https://karenwky-diabetes-predictor.hf.space/docs

• FastAPI - Modern, fast web framework for building APIs
• scikit-learn - Machine learning library for model training and prediction
• Pydantic - Data validation and settings management
• NumPy - Numerical computing library
• Uvicorn - ASGI server for running the FastAPI application
• Docker - Containerization for easy deployment
• Python 3.13 - Programming language

diabetes-predictor/
├── app/
│   ├── __init__.py
│   └── main.py              # FastAPI application and endpoints
├── models/
│   └── diabetes_model.pkl   # Trained machine learning model
├── requirements.txt         # Python dependencies
├── train_model.py           # Model training script
├── train_model.ipynb        # Jupyter notebook for model development
├── Dockerfile               # Container configuration
├── app.log                  # Model training log
└── image_info.txt           # Docker images info

1. Clone and navigate to the project:

   cd /Users/karen/Documents/aie-journey/study/deployment/diabetes-predictor

2. Install dependencies:

   pip install -r requirements.txt

3. Run the API server:

   uvicorn app.main:app --reload --host 0.0.0.0 --port 8000

4. Access the API:

   • API: http://localhost:8000
   • Interactive docs: http://localhost:8000/docs
   • Alternative docs: http://localhost:8000/redoc

1. Build the Docker image:

   docker build -t diabetes-predictor .

2. Run the container:

   docker run -p 8000:8000 diabetes-predictor

Check if the API is running:

curl http://localhost:8000/

Response:

{
  "status": "healthy",
  "model": "diabetes_progression_v1"
}

Send patient data to get diabetes progression predictions:

curl -X POST "http://localhost:8000/predict" \
     -H "Content-Type: application/json" \
     -d '{
       "age": 0.05,
       "sex": 0.05,
       "bmi": 0.06,
       "bp": 0.02,
       "s1": -0.04,
       "s2": -0.04,
       "s3": -0.02,
       "s4": -0.01,
       "s5": 0.01,
       "s6": 0.02
     }'

Response:

{
  "predicted_progression_score": 213.34,
  "interpretation": "Above average progression"
}

The API expects normalized physiological features:

• age: Patient age
• sex: Patient sex
• bmi: Body mass index
• bp: Average blood pressure
• s1: Serum measurement 1
• s2: Serum measurement 2
• s3: Serum measurement 3
• s4: Serum measurement 4
• s5: Serum measurement 5
• s6: Serum measurement 6

• < 100: Below average progression
• 100-150: Average progression
• > 150: Above average progression

import requests

# API endpoint
url = "http://localhost:8000/predict"

# Patient data
patient_data = {
    "age": 0.05,
    "sex": 0.05,
    "bmi": 0.06,
    "bp": 0.02,
    "s1": -0.04,
    "s2": -0.04,
    "s3": -0.02,
    "s4": -0.01,
    "s5": 0.01,
    "s6": 0.02
}

# Make prediction
response = requests.post(url, json=patient_data)
result = response.json()

print(f"Predicted Score: {result['predicted_progression_score']}")
print(f"Interpretation: {result['interpretation']}")

• All input features should be normalized as per the diabetes dataset standard
• The model returns a continuous progression score
• Higher scores indicate greater diabetes progression
• The API includes automatic data validation and error handling

• Step-by-Step Guide to Deploying Machine Learning Models with FastAPI and Docker - MachineLearningMastery.com
• Docker Spaces | Hugging Face Hub Documentation
• Deploying Your FastAPI Applications on Huggingface Via Docker
• Say Goodbye to Print(): Use Logging Module for Effective Debugging - KDnuggets

This project is licensed under the MIT License - see the LICENSE file for details.