IBM Research and Sustainable Energy for All (SEforALL) collaborated through the IBM Sustainability Accelerator to conceptualize and design an AI solution to predict where cities will grow, called the Modeling Urban Growth (MUG) AI model. The MUG model is trained on, and validates, historical data from satellite images; geographic data, such as slope and elevation; demographic data; and structural data, such as road layout, combining the data into a time series. MUG helps users to map future urbanization and associated infrastructure needs, enabling decision makers to prioritize communities and developing regions that need support for issues like electrification and energy services.
MUG is publicly available and open-sourced on GitHub. The code supports model training on data available for different countries.
The model is designed to be re-trained by users for any country in the world, using publicly accessible data. On GitHub, MUG includes an explanatory guidebook on running the code and making predictions using the same or different datasets, which further expands access to developers and decision-makers.
For more on MUG and its use by Sustainable Energy for All (SEforALL), see the following report