kabila96/Crop-Yield-Assessment-Dashboard

A deployable portfolio dashboard for crop production prediction using a Zimbabwe-focused synthetic dataset. # Crop Yield Assessment Dashboard **Created By Powell Andile Ndlovu** This project demonstrates a full machine learning workflow for predicting crop production from synthetic Zimbabwe-focused agricultural data. The package includes data generation, exploratory analysis, Random Forest and XGBoost modelling, explainability using SHAP, permutation importance, and LIME, plus a deployable Streamlit dashboard. ## Disclaimer This project uses a synthetic dataset created for demonstration, machine learning experimentation, and portfolio development. It is not an official agricultural statistics source. ## Files - `zimbabwe_crop_yield_bulletproof.csv` — synthetic dataset - `train_crop_yield_models.py` — training pipeline - `streamlit_app.py` — deployable dashboard - `artifacts/` — saved models and metrics - `plots/` — saved figures - `dissertation_writeup.md` — dissertation-style write-up ## Run locally ```bash pip install -r requirements.txt python train_crop_yield_models.py streamlit run streamlit_app.py ```