alhassan-taha/egypt-power-forecasting

# ⚡ Egypt Power Load Forecasting System Intelligent hourly electricity demand prediction across all governorates A Machine Learning system utilizing XGBoost, Random Forest, and Linear Regression to forecast electricity consumption in Egypt based on real-time weather data. --- ## 📌 Project Overview This project predicts the hourly electricity load (in Megawatts) for 27 Egyptian governorates. It features a professional, interactive web dashboard built with Streamlit that automatically fetches real weather data via Open-Meteo API and applies a trained Multi-Model Machine Learning pipeline to generate accurate forecasts up to 7 days ahead. ## 🚀 Features - **Multi-Model Engine:** Compare live forecasts between Random Forest, XGBoost, and Linear Regression. - **Dynamic Switching:** Instantly switch between models and see confidence margins (MAE & R2). - **Real-time Weather:** 100% real hourly weather data (Temperature, Humidity, Wind Speed) fetched automatically. - **Data Profiling:** In-depth interactive dataset dictionary, feature visualization, and descriptive statistics. - **Professional UI:** Light-themed, fully responsive dashboard with Material Icons and interactive Plotly charts. ## 🛠️ Prerequisites - Python 3.11+ - Git ## ⚙️ Installation & Setup (How to Run from Scratch) Follow these exact steps to run the project on any new machine: ### 1. Clone the Repository ```bash git clone github.com cd egypt-power-forecasting ``` ### 2. Install Dependencies Open your terminal/command prompt and run: ```bash pip install pandas numpy xgboost scikit-learn joblib streamlit plotly requests ``` ### 3. Train the Models (Generate the Pipelines) Before starting the web app, you must train the AI models. This script will train Linear Regression, Random Forest, and XGBoost on the dataset and save them into `model_pipeline.pkl`. ```bash python scratch/train_models.py ``` *(Note: Wait until you see "✅ Pipeline saved successf …