edeki101/MSc-Renewable-Forecasting

LightGBM framework for 24-hour ahead solar and wind power forecasting in Nigeria with Explainable AI (SHAP). # 🌞 MSc Renewable Energy Forecasting ## 📌 Project Overview 24-hour ahead solar and wind power forecasting for Nigeria using LightGBM with Explainable AI (SHAP) for grid operator transparency and reliability. --- ## 🎯 Problem Statement Nigeria's power grid faces significant challenges with: - Unreliable renewable energy integration - Lack of accurate forecasting tools - Poor grid operator decision making - Limited transparency in ML predictions --- ## 💡 Solution Built a LightGBM forecasting model that: - Predicts solar & wind power 24 hours ahead - Explains predictions using SHAP values - Helps grid operators make better decisions - Improves renewable energy reliability --- ## 🛠️ Tools & Technologies - **Python** — Core programming language - **LightGBM** — Fast gradient boosting ML model - **SHAP** — Explainable AI for model transparency - **Pandas** — Data manipulation - **Matplotlib/Seaborn** — Visualization - **Scikit-learn** — Model evaluation --- ## 📊 Model Performance - **Model:** LightGBM (replaced CNN-LSTM for speed) - **Target:** 24-hour ahead power forecasting - **Features:** Weather data, historical power output - **Explainability:** SHAP values for each prediction --- ## 🌍 Domain - **Field:** Electrical Engineering & Data Science - **Application:** Nigerian Power Grid - **Impact:** Improved renewable energy forecasting --- ## 👤 Author **edeki monday Ekundayo** MSc Electrical Engineer | Data Scientist - GitHub: github.com - Available for remote opportunities