---
title: Smart Fertilizer Recommender API
emoji: 🌾
colorFrom: green  
colorTo: blue
sdk: docker
pinned: false
license: mit
short_description: AI-powered fertilizer recommendations for crops
---

# 🌾 Smart Fertilizer Recommender API

A FastAPI-based REST API that provides intelligent fertilizer recommendations using a Random Forest machine learning model. The system analyzes soil nutrient parameters and crop type to suggest the most suitable fertilizer.

## 🚀 Features

- **Intelligent Recommendations**: ML-powered fertilizer suggestions based on soil analysis
- **Multi-Crop Support**: Recommendations for various crop types
- **Confidence Scoring**: Each recommendation comes with a confidence percentage
- **Input Validation**: Automatic validation of soil parameter ranges
- **Batch Processing**: Process multiple soil samples at once
- **Interactive Documentation**: Built-in Swagger UI for easy testing

## 🔬 Model Information

- **Algorithm**: Random Forest Classifier
- **Features**: Nitrogen %, Phosphorus (Olsen) ppm, Potassium meq%, Soil pH, Crop Type
- **Training Accuracy**: ~95%+
- **Cross-Validation**: 5-fold CV for robust evaluation

## 📊 API Endpoints

### Core Endpoints

- `GET /` - Welcome message and API information
- `GET /health` - Health check endpoint
- `GET /model-info` - Model performance metrics and details
- `GET /crops` - List of supported crops
- `GET /fertilizers` - List of available fertilizers

### Prediction Endpoints

- `POST /predict` - Single fertilizer recommendation
- `POST /batch-predict` - Batch processing for multiple samples

### Utility Endpoints

- `GET /docs` - Interactive Swagger documentation
- `GET /redoc` - Alternative API documentation
- `GET /debug/files` - Debug information

## 🔧 Usage Examples

### Single Prediction

```python
import requests

# API endpoint
url = "https://OwinoGoddy.hf.space/predict"

# Soil parameters
data = {
    "nitrogen_percent": 0.08,
    "phosphorus_ppm": 23.0,
    "soil_ph": 5.76,
    "potassium_meq_percent": 1.36,
    "crop": "maize"
}

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

print(f"Recommended Fertilizer: {result['primary_recommendation']}")
print(f"Confidence: {result['confidence']}%")
Response Example
json{
  "primary_recommendation": "NPK 20-10-10",
  "confidence": 87.5,
  "all_recommendations": [
    {
      "fertilizer": "NPK 20-10-10",
      "confidence": 87.5
    },
    {
      "fertilizer": "DAP",
      "confidence": 8.2
    },
    {
      "fertilizer": "Urea",
      "confidence": 3.1
    }
  ],
  "input_parameters": {
    "nitrogen_percent": 0.08,
    "phosphorus_ppm": 23.0,
    "soil_ph": 5.76,
    "potassium_meq_percent": 1.36,
    "crop": "maize"
  },
  "validation_warnings": []
}
📝 Input Parameters
ParameterTypeRangeDescriptionnitrogen_percentfloat0-5%Total Nitrogen percentagephosphorus_ppmfloat0-200 ppmPhosphorus Olsen in ppmsoil_phfloat3-10Soil pH levelpotassium_meq_percentfloat0-10%Potassium in meq%cropstring-Crop name (see /crops endpoint)
🌱 Supported Crops
The system supports various crop types including:

Maize
Beans
Wheat
Rice
Sorghum
And many more...

Use the /crops endpoint to get the complete list of supported crops.
🔍 Model Validation
The API includes automatic input validation:

Range checking for all numerical parameters
Statistical validation against training data
Crop name validation
Warning system for unusual values

🚀 Getting Started

Access the API: Visit the Hugging Face Space URL
Interactive Testing: Use /docs for Swagger UI
Get Crops: Check /crops for supported crop types
Make Predictions: Send POST requests to /predict

📈 Performance Metrics

Accuracy: High accuracy on test data
Speed: Fast predictions (<100ms typical)
Reliability: Robust error handling and validation
Scalability: Supports batch processing

🛠️ Technical Stack

Backend: FastAPI
ML Framework: scikit-learn
Model: Random Forest Classifier
Validation: Pydantic
Documentation: Swagger/OpenAPI
Deployment: Docker + Hugging Face Spaces

📞 Support
For questions, issues, or feature requests, please visit the Hugging Face Space discussion tab.
📄 License
MIT License - feel free to use this API for your agricultural applications!

Built with ❤️ for sustainable agriculture