The presentation outlines a research agenda in urban data science for just and sustainable cities, centered on spatial accessibility as a human-centered framework linking land use, public services, and transportation planning. It conceptualizes spatial access to opportunities as fundamental to meeting basic needs, reducing inequalities, supporting human development, and shaping environmentally sustainable mobility patterns. The talk presents empirical research showing how access to opportunities is a key concept to understand pressing challenges of urban growth, spatial segregation and social inclusion. It includes a detailed case study on welfare policies in Brazil, presenting causal evidence that improving spatial proximity to social assistance reference centers helps increase welfare take-up among vulnerable populations. Methodologically, it highlights recent developments in spatial data science and open-source computational tools to support policy-relevant urban analysis. Finally, it outlines future research directions that move from inequalities in access opportunities to inequalities in socioeconomic and health outcomes, emphasizing causal inference, policy evaluation, and the use of spatial data science to inform more equitable and sustainable urban planning.