| Süleyman Nurullah Adahi Şahin, Abdulkadir Özden, Hakan Kocaman, Abdurrahman Korkmaz, Yavuz Halimergün, Cem Özkurt Mapping Mobility Justice: A Deep Learning and GIS-Based Framework for Micro-Scale Sustainable Mobility Policy Assessment |
|---|
| Abstract. The development of inclusive and sustainable transport policies demands ac-curate, spatially detailed data on the built environment and mobility infra-structure—particularly at the neighborhood level, where daily mobility pat-terns and access disparities are most visible. This study proposes MiSHar (Micro-Scale Sustainable Mobility Index), an integrated analytical framework that leverages deep learning-based object detection and Geographic Information Systems (GIS) to systematically evaluate the sustainable mobility potential of urban micro-regions. High-resolution satellite and aerial imagery are processed using convolutional neural networks to automatically detect physical features such as buildings, green areas, paved surfaces, pedestrian and bicycle networks. These are subsequently combined with GIS-based public transport bus stop yielding a multi-dimensional index formulation. The index incorporates weighted parameters related to population density, infrastructure provision, and land use patterns. A pilot implementation in Adapazarı District (Sakarya, Türkiye) demonstrates the model’s ability to re-veal intra-urban mobility inequalities and provide spatial evidence for transport-oriented development strategies. MiSHar represents a policy-responsive, survey-independent, and scalable approach to mobility diagnos-tics. By integrating AI-powered spatial analytics with urban planning princi-ples, it offers a valuable tool for local governments seeking to target infra-structure investments, design equity-focused interventions, and align with European and global sustainable mobility goals. |
| Keywords: Sustainable mobility index, Deep learning, GIS, Micro-scale analysis |
 Download PDF |
| DOI: https://doi.org/10.54381/itta2026.3.06 |