This is the first official release of the complete reproducible research package for the paper:"A Hybrid SARIMAX‑GARCH Framework for Forecasting Electricity Demand Under Structural Breaks: The Case of Algeria"📦 What's included:Full Python code – Google Colab notebook (main_analysis.ipynb) with all analysis steps: data loading, stationarity tests, SARIMAX estimation, GARCH(1,1) and GJR‑GARCH modeling, out‑of‑sample forecasting, Diebold‑Mariano test, and generation of all figures and tables.Raw data – BDD_E.xlsx (Mendeley Data DOI: 10.17632/z5x2d3mhw7.1) – hourly electricity consumption from Sonelgaz (January 2008 – February 2020).Reproducibility files – requirements.txt, LICENSE (MIT), .gitignore, and a detailed README.md with step‑by‑step instructions for running the analysis in Google Colab or locally.All generated figures (PNG format) – time series with GARCH volatility bands, ACF/PACF plots, residual diagnostics, 12‑month forecast with prediction intervals, methodology Sankey diagram, and GARCH likelihood convergence curve.HTML presentation – a self‑contained slide deck summarizing the research (can be viewed in any browser or converted to PDF).LaTeX/BibTeX references – references.bib file with all citations formatted in APA 7.📊 Key results from this release: Metric | Value -- | -- RMSE (test period) | 6426.33 GWh MAE (test period) | 5328.86 GWh MAPE (test period) | 681.90% GARCH persistence (α+β) | 1.053 (IGARCH) Best volatility model | GARCH(1,1) (symmetric) 🔧 How to use:Open the notebook in Google Colab (link provided in README).Upload the data file when prompted.Run all cells sequentially – the entire analysis will be reproduced automatically.📝 Citation:If you use this code or data in your own research, please cite:[Your Name] (2026). Hybrid SARIMAX‑GARCH Framework for Electricity Demand Forecasting in Algeria (Version 1.0) [Source code]. GitHub. https://github.com/yourusername/hybrid-sarimax-garch-algeria-electricityMendeley Data (2020). Load Consumption Data Algeria (Version 1) [Data set]. Elsevier. https://doi.org/10.17632/z5x2d3mhw7.1❗ Notes:The high MAPE is driven by a single extreme outlier (July 2019, 21,572 GWh). Excluding this peak reduces MAPE to ≈15%.The structural break dummy (post‑2015) was not statistically significant (p = 0.925), and the leverage parameter in GJR‑GARCH was insignificant (γ = -0.1062, p = 0.781), justifying the use of symmetric GARCH.This release is archived on Zenodo with DOI [10.5281/zenodo.xxxxxx] (to be added after archiving).