pryyyynz/Ghana-Indigenous-Challenge-Zindi

# Ghana Indigenous Weather Forecasting — Rainfall Prediction A friendly, high-level overview of this repository to help you orient quickly. This README focuses on what’s here and how it’s organized, not on running code. For execution details, see the QUICKSTART/README files inside each implementation folder. ## What This Repo Contains - A supervised learning project for daily rainfall prediction, built around multiple incremental implementations (v1 → v4) exploring preprocessing, modeling (notably CatBoost), interpretability, and deployment artifacts. - A starter notebook for exploratory work, plus CSV datasets provided by the challenge. - Generated outputs (models, submissions, interpretability reports) saved per implementation. ## Top-Level Files - Indigenous_Weather_Forecasting-Starter_Notebook.ipynb — quick EDA/prototyping entry point. - train.csv, test.csv — challenge datasets. - SampleSubmission.csv — required output format sample. - manifest-*.json — environment/artifact metadata (name may vary). ## Implementations (Folders) Each implementation is a self-contained take on the pipeline. They follow a similar structure and evolve in complexity and features. - rainfall_prediction/ - Early pipeline with `main.py` and `train_pipeline.py`, plus `deployment.py` and `interpretability.py`. - Requirements: requirements_core.txt and requirements.txt. - Outputs in rainfall_prediction/outputs/ and rainfall_prediction/outputs_/. - rainfall_prediction_2/ - Refined structure with `train.py`, `config.py`, `models.py`, `preprocessing.py`, `deployment.py`, and `interpretability.py`. - Example inference script: `example_inference.py`. - Outputs in rainfall_prediction_2/outputs/ and rainfall_prediction_2/outputs_/. - rainfall_prediction_3/ - Adds `resampling.py` and additional reports/notes (e.g., ARCHITECTURE.md, PROJECT_SUMMARY.md). - Similar module layout: `train.py`, `config.py`, `models.py`, `preprocessing.py`, `interpretability.py`, `deployment.py`. - rainfall_predi …