About Compact City AI
Compact City AI uses machine learning to evaluate urban livability by analyzing spatial access to amenities within walkable distances around transit stations. By leveraging neural networks, graph-based models, and real-time data, it predicts station-level scores to support compact, sustainable city planning.
Compact City AI is an AI-driven platform that assesses livability in urban areas by analyzing accessibility to essential services—such as schools, healthcare, restaurants, parks, and fitness centers—within a 600-meter radius of transit stations. The project leverages multiple machine learning techniques: Artificial Neural Networks (ANNs) model nonlinear patterns in urban service availability; Convolutional Neural Networks (CNNs) detect spatial density and clustering of amenities; and Graph Neural Networks (GNNs) capture the connectivity and influence of nearby neighborhoods and infrastructure. Regression models are used to predict station-level livability scores based on feature-weighted proximity and service diversity. All models are developed in Python and integrated into KNIME workflows for visual, modular development and deployment. The system operates in real-time using APIs like Google Maps and OpenStreetMap, enabling users to input any city and instantly receive AI-powered livability maps and station rankings. Compact City AI supports urban planners, researchers, and citizens in making data-informed decisions toward more walkable, inclusive, and efficient urban environments.
The Team
Our team at Compact City AI brings together professionals from Japan, India, and the United States, combining deep expertise in urban planning, artificial intelligence, and geospatial analytics. This international collaboration enables us to address smart city challenges with globally informed strategies and locally adaptable solutions. The project is supported by academic partnerships between the Faculty of Urban Science at Meijo University (Japan) and the School of Planning and Architecture, New Delhi (India).Together, we co-develop AI-powered tools and real-time urban intelligence platforms that promote compact, sustainable, and livable cities. The Faculty of Urban Science at Meijo University, located in Nagoya, Japan, plays a pioneering role in interdisciplinary urban research, focusing on smart cities, public participation, and sustainable design. The School of Planning and Architecture, New Delhi, brings decades of expertise in urban and regional planning, offering deep contextual knowledge for emerging urban centers. Complementing these efforts, the Spatial Data Lab at Harvard University contributes advanced capabilities in spatial computing, data visualization, and global urban analytics to machine learning platforms like KNIME. This cross-continental partnership strengthens our mission to create scalable, AI-driven solutions for compact city development that serve governments, citizens, and businesses alike.