Cities continuously generate massive, heterogeneous, and multidimensional data, yet their potential remains underexploited without semantically coherent models, reproducible workflows, and sustainable computational infrastructures. The central concept of semantic lenses captures a methodological innovation: the ability to focus on specific disciplinary perspectives, defocus to integrate multiple viewpoints, and refocus as urban narratives evolve—an essential capability for shaping cities that are not only smart and sustainable but also inclusive and adaptive to diverse urban realities.

A major contribution of this work lies in the development of unified frameworks for representing and interlinking urban data through knowledge graphs, 3D city models, and nD (spatial, temporal, thematic, semantic,\dots) urban digital twins. The proposed methodologies and their associated data infrastructures enable the fusion of multisource, multiscale datasets while simultaneously ensuring provenance, interoperability, and semantic richness—all key prerequisites for intelligent, data-driven urban management. The semantic lenses framework, operationalized through versioned and concurrent knowledge graphs, allows dynamic refocusing between detailed disciplinary views and broader integrative perspectives, thus facilitating transparent, reproducible, and context-aware decision-making.

Building upon these semantic foundations, my research strengthens the reproducibility of methodological contributions through provenance-aware workflows and benchmarked algorithms specifically designed for large-scale and evolving urban datasets. The sustainability dimension guides ongoing efforts to optimize computational efficiency, scalability, and environmental responsibility, while explicitly addressing the energy costs of data-intensive analytics. In parallel, the inclusivity dimension is reinforced through multimodal visualizations, multilingual knowledge bases, and accessible visual interfaces that democratize access to urban knowledge and empower citizens, researchers, and institutions alike.

Beyond industrial adoption, my research has also been applied to the virtual reconstruction and visualization of the voussoirs of Notre-Dame de Paris, as well as to ongoing projects focused on urban sustainability initiatives.Beyond scientific contributions, the HDR thesis highlights the societal and educational impact of urban data science as a maturing, interdisciplinary field. A decade of mentorship and teaching has integrated computer science, data science, and artificial intelligence with openness, critical thinking, and lifelong learning. Dissemination of contributions through open-source software, collaborative platforms, and partnerships with associations and cultural institutions demonstrates how semantic lenses foster participatory innovation, and inclusive smart city ecosystems.

This habilitation thesis thus presents a coherent trajectory showing how urban data science can evolve into a reproducible, interdisciplinary, and socially responsive discipline grounded in semantic web technologies. By aligning computational rigor with sustainability, intelligence, and inclusivity, it lays the foundations for transparent and adaptive infrastructures that serve the complex and diverse realities of contemporary cities.