Origin-Destination (OD) flow matrices are critical for urban mobility analysis, supporting traffic forecasting, infrastructure planning, and policy design. Existing methods face two key limitations: (1) reliance on costly auxiliary features (e.g., Points of Interest, socioeconomic statistics) with limited spatial coverage, and (2) fragility to spatial topology changes, where reordering urban regions disrupts the structural coherence of generated flows. We propose Sat2Flow, a structure-aware diffusion framework that generates structurally coherent OD flows using only satellite imagery. Our approach employs a multi-kernel encoder to capture diverse regional interactions and a permutation-aware diffusion process that maintains consistency across regional orderings. Through joint contrastive training linking satellite features with OD patterns and equivariant diffusion training enforcing structural invariance, Sat2Flow ensures topological robustness under arbitrary regional reindexing. Experiments on real-world datasets show that Sat2Flow outperforms physics-based and data-driven baselines in accuracy while preserving flow distributions and spatial structures under index permutations. Sat2Flow offers a globally scalable solution for OD flow generation in data-scarce environments, eliminating region-specific auxiliary data dependencies while maintaining structural robustness for reliable mobility modeling.

Accurate forecasting of granular socioeconomic indicators, such as GDP, is essential for informed economic decision-making. Despite pioneering efforts to harness multi-modal data using traditional supervised or self-supervised learning methods for economic prediction, effectively integrating their features remains a significant hurdle. To address the challenge, we propose a semi-supervised enhanced graph learning framework, SemiGPS, which leverages multi-modal geospatial big data to extract features linked to regional economic development. By integrating multiple data modalities, such as street view images and POIs, and adopting a semi-supervised learning paradigm, SemiGPS strikes a balance between spatial-interacted learning from self-supervision and the fitting capabilities from supervised learning, thereby empowering the model to learn more informative and effective representations. Experiments in the Pearl River Delta region of China demonstrate the effectiveness of our approach, with R2scores of 0.85, 0.95, and 0.88 for the primary, secondary, and tertiary sectors, respectively. Our results highlight the potential of SemiGPS to capture nuanced regional economics and pave the way for more accurate economic forecasting using geospatial data.