SEED

A comprehensive system for agricultural data generation and machine learning predictions for The Gambia.

!NOTE

this is a learning project for Dawda Borje Kujabi to learn machine learning

🌾 Features

Data Generation

• Historical Data: Agricultural datasets from 2001-2021
• Real Data Integration: Uses actual datasets from data/ directory
• Multiple Categories: Crops, Fisheries, Sales, Livestock, Farm Practices, Land Tenure
• Growth Projections: Realistic growth patterns with climate variability

Machine Learning Model

• Multi-Target Prediction: Yield, Price, and Production
• Environmental Factors: Rainfall, temperature, humidity, soil pH
• Economic Factors: Fuel prices, labor costs, market demand
• Technology Factors: Fertilizer use, irrigation, crop type
• Future Predictions: Support for current and future years (2024-2030+)

🚀 Quick Start

Run Main Predictions

# Run comprehensive predictions for current and future years
python seed/main.py

Run Specific Year Predictions

# Predict for current year
python scripts/run_predictions.py 2024

# Predict for future year
python scripts/run_predictions.py 2029

# Predict specific crop for 2025
python scripts/run_predictions.py 2025 --crop Rice

# Rainfall impact analysis for 2026
python scripts/run_predictions.py 2026 --rainfall

Test Model

# Simple model test
python test_model.py

📊 Prediction Capabilities

Current Year Predictions

• 2024: Current year with real data
• Good/Average/Poor Conditions: Multiple scenarios
• All Crops: Rice, Millet, Sorghum, Maize, Groundnuts, Cotton, Vegetables, Fruits

Future Year Predictions

• 2025-2030+: Future predictions with trend adjustments
• Climate Change: Temperature and rainfall projections
• Technology Adoption: Increasing fertilizer and irrigation use
• Economic Trends: Rising fuel prices and labor costs
• Market Evolution: Growing demand for certain crops

Rainfall Impact Analysis

• Rainfall Levels: 400-1400mm range
• Crop Sensitivity: Different crops respond differently to rainfall
• Yield Impact: Shows how rainfall affects crop yields
• Price Effects: Rainfall impact on crop prices

📁 Project Structure

seed/
├── data/                    # Agricultural datasets (2001-2021)
│   ├── 2001/              # Year-specific data
│   ├── 2002/
│   └── ...                # Up to 2021
├── scripts/                # Data generation scripts
│   ├── create_datasets.py
│   ├── run_predictions.py
│   └── README.md
├── seed/
│   ├── model.py           # ML model using real data
│   └── main.py            # Main prediction application
├── external_data/         # Excel data source
└── test_model.py          # Simple model test

🎯 Model Features

Input Variables

• Rainfall (mm/year): Primary environmental factor
• Temperature (°C): Climate impact
• Humidity (%): Moisture conditions
• Soil pH: Soil quality indicator
• Fertilizer Use (kg/ha): Technology adoption
• Irrigation Area (%): Water management
• Fuel Price (USD/liter): Economic factor
• Labor Cost (USD/day): Economic factor
• Market Demand (index): Market conditions
• Crop Type: Categorical variable

Output Predictions

• Yield (tons/hectare): Crop productivity
• Price (USD/ton): Market value
• Production (tons): Total output

Machine Learning Models

• Random Forest: Best for complex relationships
• Gradient Boosting: Handles non-linear patterns
• Ridge Regression: Linear relationships
• Linear Regression: Baseline model

📈 Future Predictions

The system can predict for any future year with realistic adjustments:

Climate Trends (2024-2030)

• Temperature: +1% annually (climate change)
• Rainfall: +0.5% annually (slight increase)
• Humidity: Stable with minor variations

Technology Trends

• Fertilizer Use: +3% annually (increasing adoption)
• Irrigation: +3% annually (water management)
• Crop Yields: +2% annually (improved practices)

Economic Trends

• Fuel Prices: +5% annually (inflation)
• Labor Costs: +5% annually (wage increases)
• Market Demand: +2% annually (population growth)

🌧️ Rainfall Impact Analysis

The model shows how different rainfall levels affect crop performance:

Rainfall (mm)	Rice Yield	Millet Yield	Groundnuts Yield
400	Low	Medium	Low
600	Medium	High	Medium
800	High	High	High
1000	Very High	High	Very High
1200	Very High	Medium	Very High
1400	Very High	Low	Very High

🎯 Usage Examples

Predict Rice Production for 2029

from seed.model import GambiaCropPredictor

model = GambiaCropPredictor()
real_data = model.load_real_data()
model.train_models(real_data)

# Predict for 2029
input_data = {
    "crop": "Rice",
    "rainfall_mm": 850,
    "temperature_c": 28.5,  # Adjusted for 2029
    "humidity_percent": 75,
    "soil_ph": 6.8,
    "fertilizer_use_kg_ha": 85,  # Higher adoption
    "irrigation_area_percent": 35,  # More irrigation
    "fuel_price_usd_liter": 1.8,  # Higher fuel prices
    "labor_cost_usd_day": 22,  # Higher labor costs
    "market_demand_index": 120,  # Higher demand
}

predictions = model.predict(input_data)
print(f"Yield: {predictions['yield'][0]:.2f} tons/hectare")
print(f"Price: ${predictions['price'][0]:.2f}/ton")
print(f"Production: {predictions['production'][0]:.0f} tons")

📋 Requirements

• Python 3.8+
• pandas
• numpy
• scikit-learn
• joblib

🔧 Installation

# Install dependencies
pip install pandas numpy scikit-learn joblib

# Run predictions
python seed/main.py

📊 Data Sources

• FAO Agricultural Census Report 2001/2002
• GBOS (Gambia Bureau of Statistics) Data
• Real datasets from data/ directory
• Environmental and economic projections

The system is now ready for comprehensive agricultural predictions for The Gambia, with support for current and future years based on real data and realistic projections.