sayedkhaledelsayed/Egypt-Crop-Yield-Prediction-99.17-XGBoost-

# Egypt-Crop-Yield-Prediction-99.17-XGBoost- # 🌾 Egypt Crop Yield Prediction (1990-2024) --- ## 🎯 Overview This project provides a data-driven analysis and prediction of agricultural crop yields in Egypt over a **34-year period (1990-2024)**. By merging geographical knowledge with advanced Machine Learning techniques, I developed a model capable of forecasting yields with high precision. > **Key Achievement:** Reached a **99.2% R² Score** using the XGBoost Regressor. "The model was validated using 10-Fold Cross-Validation with shuffling, achieving a consistent mean R2 score of 99.12%, ensuring its reliability across different data samples." ## 📊 The Dataset The data is sourced from the **Food and Agriculture Organization (FAOSTAT)**. - **Raw Data:** Comprehensive records of Egyptian agriculture. - **Processing:** Data was cleaned, and features were engineered using **Pivot Tables** to align 'Area Harvested', 'Production', and 'Yield' for each crop type per year. ## 🛠️ Tech Stack & Tools - **Language:** Python 🐍 - **Libraries:** Pandas, NumPy, Scikit-learn, XGBoost. - **Visualization:** Matplotlib, Seaborn. - **Environment:** Kaggle Notebooks. ## 🚀 Project Workflow 1. **Data Cleaning:** Handling missing values and filtering for Egypt-specific records. 2. **Feature Engineering:** Reshaping data from long-form to wide-form (Pivoting). 3. **Model Selection:** Comparing different regressors (Random Forest vs. XGBoost). 4. **Hyperparameter Tuning:** Optimizing XGBoost for maximum accuracy. 5. **Serialization:** Saving the final model and encoders using `joblib` for deployment. ## 📂 Repository Structure - `final_crop_model.pkl`: The trained XGBoost model. - `crop_encoder.pkl`: Label encoder for crop types. - `egypt_crop_yield_processed_pivot.csv`: The cleaned dataset used for training. ## 💻 Usage To use the model in your environment: ```python import joblib import pandas as pd # Load model and encoder model = joblib.load('final_crop_model.pkl') encoder = jo …