LidetuTadesse/Suitable_crop_recommendation_and_Yield_predition

An end-to-end machine learning project for crop suitability recommendation and yield prediction in Ethiopia. It integrates soil, climate, and agronomic data, applies robust preprocessing and feature engineering, trains classification and regression models, and produces explainable, reproducible results using Git, DVC, and Docker. # Crop Suitability and Yield Prediction Using Machine Learning ## Project Overview This project is an end-to-end machine learning system designed to support crop suitability recommendation and yield prediction using soil, climate, and agronomic data. The primary motivation is to improve data-driven agricultural decision-making, with a particular focus on Ethiopian cereal crops. The project follows a research-oriented yet production-ready structure, emphasizing reproducibility, explainability, and clear separation between experimentation and deployable code. --- ## Objectives - Analyze soil and climate conditions relevant to major cereal crops - Handle agronomic data expressed as ranges (e.g., nutrient levels, rainfall, temperature) - Develop machine learning models for: - Crop suitability recommendation - Crop species classification - Yield prediction - Provide interpretable model outputs to support agronomic insights - Ensure full reproducibility using modern MLOps practices --- ## Data Description The dataset includes soil, climate, and agronomic variables such as: - Soil nutrients (Nitrogen, Phosphorus, Potassium) - Soil pH - Temperature and rainfall - Altitude and length of growing period (LGP) - Crop type and crop species - Yield ranges Many variables are represented as ranges derived from agronomic studies. These ranges are explicitly handled during preprocessing and feature engineering rather than being treated as raw numeric values. --- ## Methodology The project follows a structured machine learning workflow: 1. Exploratory Data Analysis (EDA) to understand distributions, relationships, and data quality 2. Preprocessing and validation of range-based agronomic data 3. Feature engineering for crop suitability and yield modeling 4. Model training using classical and ensemble machine learning methods 5. Model evaluation and explainability analysis 6. Optional deployment through an API for practical use --- ## Project Structure The repository is organ …