Multi-view region embedding has become an important technique to be pursued due to the increasing availability of diverse urban sensing data. Existing methods typically adopt attention-based fusion or contrastive alignment to integrate multiple data sources such as human mobility and points-of-interest (POIs) into unified region representations. However, these approaches often fail to effectively balance cross-view commonality modeling with the preservation of view-specific characteristics. To address this limitation, we propose ComSRE, a novel region embedding framework that explicitly disentangles shared and distinctive components in multi-view region representations. ComSRE integrates three key modules: (1) a Multi-view Representation Learning module that fuses complementary information across views, (2) a View-to-Commonality Contrastive Alignment module that aligns the representation of each view to a shared commonality anchor to enhance cross-view consistency, and (3) a Multi-view Differential Orthogonality module that isolates distinctive signals unique to each view and promotes their independence through orthogonality constraints. Extensive experiments on three real-world urban datasets demonstrate that ComSRE consistently outperforms state-of-the-art methods across multiple downstream tasks, achieving superior predictive accuracy and representation quality.

Predicting human mobility remains a fundamental challenge, especially when individuals deviate from routine patterns due to exploration, disruptions, or rare events. While sequential models like Transformers excel at capturing regular movement patterns, their performance often degrades under nonroutine scenarios involving rare or unfamiliar transitions. To address this, we propose ROAM (Routine-Oriented Adaptive Mobility Predictor), a novel framework that jointly models human mobility from both sequential and relational perspectives, enabling adaptive handling of both routine and nonroutine behaviors during prediction. ROAM combines a sequential encoder that captures historically frequent transitions with a complementary graph-based relational reasoning module that encodes both user-specific and group-level mobility structures. To dynamically integrate these views, we introduce a hierarchical confidence-aware gating mechanism that adaptively balances sequential and relational predictions based on their internal reliability. Extensive experiments on real-world mobility datasets show that ROAM consistently outperforms state-of-the-art baselines in next location prediction. Further analysis reveals that the combination of sequential and relational reasoning substantially improves robustness, particularly under out-of-routine scenarios.

Region representation learning emerges as a new research paradigm to encode urban systems and facilitate geographic mapping. Recent studies have sought to reasonably introduce inductive biases, which refer to prior assumptions that guide model learning, from a geospatial perspective to improve the quality of region representations. However, there remain challenges in incorporating the spatial effects, e.g. spatial dependency and spatial heterogeneity, into inductive biases, as they are critical to the geographic awareness of region representations. In response, we developed a novel region representation learning framework, termed Region Graph Spatial Contrastive Learning (RGSCL), by leveraging building footprints and points of interest (POIs) along with prior spatial knowledge to derive region representations. Specifically, RGSCL first constructed multi-view region graphs with POIs, building footprints and their spatial proximity, to form a base representation space. Next, the algorithm adopted a contrastive learning mechanism with spatial effects to formulate a dual-spatial-contrast loss function to optimise the representation space. The dual-spatial-contrasts captured POI-building spatial dependency and the region’s spatial heterogeneity to compose semantics in region representations. Experimental results demonstrated that RGSCL improved performance in geographic mapping. This study offers new insights into GeoAI from the perspective of inductive biases with respect to spatial effects.

A recent trend in urban computing involves utilizing multi-modal data for urban region embedding, which can be further expanded in a variety of downstream urban sensing tasks. Many previous studies rely on multi-graph embedding techniques and follow a two-stage paradigm: first building a k-nearest neighbor graph based on fixed region correlations for each view, and then blending multi-view information in a posterior stage to learn region representations. However, multi-graph construction and multi-graph representation learning are not associated in most existing two-stage studies, and the relationship between them is not leveraged, which can provide complementary information to each other. In this paper, we unify these two stages into one by constructing learnable weighted complete graphs of regions and propose a new one-stage Region Embedding method with Adaptive region correlation Discovery (READ). Specifically, READ comprises three modules, including a disentangled region feature learning module utilizing a city-context Transformer to encode regions' semantic and mobility features, and an adaptive weighted multi-graph construction module that builds multiple complete graphs with learnable weights based on disentangled features of regions. In addition, we propose a multi-graph representation learning module to yield effective region representations that integrate information from multiple graphs. We conduct thorough experiments on three downstream tasks to assess READ. Experimental results demonstrate that READ considerably outperforms state-of-the-art baseline methods in urban region embedding.

As geospatial data from web platforms becomes increasingly accessible and regularly updated, urban representation learning has emerged as a critical research area for advancing urban planning. Recent studies have developed foundation model-based algorithms to leverage this data for various urban-related downstream tasks. However, current research has inadequately explored deep integration strategies for multiscale, multimodal urban data in the context of urban foundation models. This gap arises primarily because the relationships between micro-scale (e.g., individual points of interest and street view imagery) and macro-scale (e.g., region-wide satellite imagery) urban features are inherently implicit and highly complex, making traditional interaction modeling insufficient. This paper introduces a novel research problem – how to learn multiscale urban representations by integrating diverse geographic data modalities and modeling complex multimodal relationships across different spatial scales. To address this significant challenge, we propose UrbanMFM, a spatial graph-based multiscale foundation model framework explicitly designed to capture and leverage these intricate relationships. UrbanMFM utilizes a self-supervised learning paradigm that integrates diverse geographic data modalities, including POI data and urban imagery, through novel contrastive learning objectives and advanced sampling techniques. By explicitly modeling spatial graphs to represent complex multiscale urban relationships, UrbanMFM effectively facilitates deep interactions between multimodal data sources. Extensive experiments on datasets from Singapore, New York, and Beijing demonstrate that UrbanMFM outperforms the strongest baselines significantly in four representative downstream tasks. By effectively modelling spatial hierarchies with diverse data, UrbanMFM provides a more comprehensive and adaptable representation of urban environments.

Predicting individuals' next locations is a core task in human mobility modelling, with wide-ranging implications for urban planning, transportation, public policy and personalised mobility services. Traditional approaches largely depend on location embeddings learned from historical mobility patterns, limiting their ability to encode explicit spatial information, integrate rich urban semantic context, and accommodate previously unseen locations. To address these challenges, we explore the application of CaLLiPer—a multi-modal representation learning framework that fuses spatial coordinates and semantic features of points of interest through contrastive learning—for location embedding in individual mobility prediction. CaLLiPer's embeddings are spatially explicit, semantically enriched, and inductive by design, enabling robust prediction performance even in scenarios involving emerging locations. Through extensive experiments on four public mobility datasets under both conventional and inductive settings, we demonstrate that CaLLiPer consistently outperforms strong baselines, particularly excelling in inductive scenarios. Our findings highlight the potential of multi-modal, inductive location embeddings to advance the capabilities of human mobility prediction systems. We also release the code and data (https://github.com/xlwang233/Into-the-Unknown) to foster reproducibility and future research.

Next location prediction is a critical task in human mobility modeling, with broad applications in personalized recommendation, urban planning, and location-based services. Recently, researchers have used prompt-based large language models (LLMs) to improve next location prediction with pre-trained knowledge. However, they face inherent challenges in bridging the gap between textual prompts for semantic contextual understanding and human mobility data for transition pattern modeling. In this paper, we introduce SILO, a framework designed for Semantic Integration in LOcation prediction via LLMs. We first construct a hybrid semantic space that seamlessly integrates ID-based embeddings, text-derived semantics, and auxiliary contextual information, enabling comprehensive modeling of sequential mobility patterns alongside contextual nuances. We then propose user-centric prompts that specify the prediction task for LLMs while embedding user context within a special token. Further, we utilize LLMs as the prediction backbone to process both user-specific prompts and hybrid ID-context embeddings of location sequences. To enhance predictive performance, we finally introduce a dual-logits strategy, combining sequential transition logits with user profile-guided semantic preference logits. Extensive experiments on two large-scale real-world mobility datasets demonstrate that SILO significantly outperforms state-of-the-art baselines, validating its effectiveness in modeling complex mobility patterns through semantic integration using LLMs.

The emerging field of geo-foundation models (GeoFM) has the potential to reshape GeoAI and spatial data science research, education, and practice. In this work, we motivate and define the term and put it into its historic context within GeoAI and spatial data science more broadly. Next, we review core datasets, models, and benchmarks. Based on this overview of the state-of-the-art, we introduce key research challenges for future GeoFM research, such as GeoAI scaling laws, geo-alignment of AI, truly multimodal GeoFM, and so on. Finally, we discuss potential risks of GeoFM research and outline the road ahead with a specific focus on the increasing role of international large-scale collaborations and the future of GeoAI and spatial data science education.

Pre-trained Foundation Models (PFMs) have ushered in a paradigm-shift in AI, due to their ability to learn general-purpose representations that can be readily employed in downstream tasks. While PFMs have been successfully adopted in various fields such as NLP and Computer Vision, their capacity in handling geospatial data remains limited. This can be attributed to the intrinsic heterogeneity of such data, which encompasses different types, including points, segments and regions, as well as multiple information modalities. The proliferation of Volunteered Geographic Information initiatives, like OpenStreetMap, unveils a promising opportunity to bridge this gap. In this paper, we present CityFM, a self-supervised framework to train a foundation model within a selected geographical area. CityFM relies solely on open data from OSM, and produces multimodal representations, incorporating spatial, visual, and textual information. We analyse the entity representations generated by our foundation models from a qualitative perspective, and conduct experiments on road, building, and region-level downstream tasks. In all the experiments, CityFM achieves performance superior to, or on par with, application-specific algorithms.

With the expansion and growth of cities, profiling urban areas with the advent of multi-modal urban datasets (e.g., points-of-interest and street view imagery) has become increasingly important in urban planing and management. Particularly, street view images have gained popularity for understanding the characteristics of urban areas due to its abundant visual information and inherent correlations with human activities. In this study, we define a street segment represented by multiple street view images as the minimum spatial unit for analysis and predict its functional and socioeconomic indicators, which presents several challenges in modeling spatial distributions of images on a street and the spatial topology (adjacency) of streets. Meanwhile, Large Language Models are capable of understanding imagery data based on its extraordinary knowledge base and unveil a remarkable opportunity for profiling streets with images. In view of the challenges and opportunity, we present a semi-supervised Urban Street Profiling Model (USPM) based on street view imagery and spatial adjacency of urban streets. Specifically, given a street with multiple images, we first employ a newly designed spatial context-based contrastive learning method to generate feature vectors of images and then apply the LSTM-based fusion method to encode multiple images on a street to yield the street visual representation; we then create the descriptions of street scenes for street view images based on the SPHINX (a large language model) and produce the street textual representation; finally, we build an urban street graph based on spatial topology (adjacency) and employ a semi-supervised graph learning algorithm to further encode the street representations for prediction. We conduct thorough experiments with real-world datasets to assess the proposed USPM. The experimental results demonstrate that USPM considerably outperforms baseline methods in two urban prediction tasks.

Next location prediction is a crucial task in human mobility modeling, and is pivotal for many downstream applications like location-based recommendation and transportation planning. Although there has been a large body of research tackling this problem, the usefulness of user preference and temporal regularity remains underrepresented. Specifically, previous studies usually neglect the explicit user preference information entailed from human trajectories and fall short in utilizing the arrival time of next location, as a key determinant on next location. To address these limitations, we propose a Multi-Context aware Location Prediction model (MCLP) to predict next locations for individuals, where it explicitly models user preference and the next arrival time as context. First, we utilize a topic model to extract user preferences for different types of locations from historical human trajectories. Second, we develop an arrival time estimator to construct a robust arrival time embedding based on the multi-head attention mechanism. The two components provide pivotal contextual information for the subsequent prediction. Finally, we utilize the Transformer architecture to mine sequential patterns and integrate multiple contextual information to predict the next locations. Experimental results on two real-world mobility datasets show that our proposed MCLP outperforms baseline methods.

Inferring urban functions using street view images (SVIs) has gained tremendous momentum. The recent prosperity of large-scale vision-language pretrained models sheds light on addressing some long-standing challenges in this regard, for example, heavy reliance on labeled samples and computing resources. In this paper, we present a novel prompting framework for enabling the pretrained vision-language model CLIP to effectively infer fine-grained urban functions with SVIs in a zero-shot manner, that is, without labeled samples and model training. The prompting framework UrbanCLIP comprises an urban taxonomy and several urban function prompt templates, in order to (1) bridge the abstract urban function categories and concrete urban object types that can be readily understood by CLIP, and (2) mitigate the interference in SVIs, for example, street-side trees and vehicles. We conduct extensive experiments to verify the effectiveness of UrbanCLIP. The results indicate that the zero-shot UrbanCLIP largely surpasses several competitive supervised baselines, e.g. a fine-tuned ResNet, and its advantages become more prominent in cross-city transfer tests. In addition, UrbanCLIP’s zero-shot performance is considerably better than the vanilla CLIP. Overall, UrbanCLIP is a simple yet effective framework for urban function inference, and showcases the potential of foundation models for geospatial applications.

The prosperity of crowdsourcing geospatial data provides increasing opportunities to understand our cities. In particular, OpenStreetMap (OSM) has become a prominent vault of geospatial data on the Web. In this context, learning urban region representations from OSM data, which is unexplored in previous work, could be profitable for various downstream tasks. In this work, we utilize OSM buildings (footprints) complemented with points of interest (POIs) to learn region representations, as buildings' shapes, spatial distributions, and properties have tight linkages to different urban functions. However, appealing as it seems, urban buildings often exhibit complex patterns to form dense or sparse areas, which brings significant challenges for unsupervised feature extraction. To address the challenges, we propose RegionDCL1, an unsupervised framework to deeply mine urban buildings. In a nutshell, we leverage random points generated by Poisson Disk Sampling to tackle data-sparse areas and utilize triplet loss with a novel adaptive margin to preserve inter-region correlations. Furthermore, we train our model with group-level and region-level contrastive learning, making it adaptive to varying region partitions. Extensive experiments in two global cities demonstrate that RegionDCL consistently outperforms the state-of-the-art counterparts across different region partitions, and outputs effective representations for inferring urban land use and population density.

Investigating the relationship between built environment factors and roadway safety is crucial for preventing road traffic accidents. Although studies have analyzed traffic-related built environment factors based on pre-determined zonal units, conclusive evidence regarding the relationship between streetscape features and traffic accidents at a fine-grained road segment level is still lacking. With the widespread availability of large-scale street view images, automatically analyzing urban built environments on a large scale is possible. Therefore, the aim of this study was to investigate the relationship between streetscape features and traffic accidents at a fine-grained road segment level using street view images. Specifically, we employed semantic image segmentation to extract streetscape elements from urban street view images, and then created traffic crash-related variables, including the street-level built environment variables, traffic variables, land-use indices, and proximity characteristics, at the road-segment level. Finally, we adopted a classification-then-regression strategy to model the number of traffic crashes while considering the zero-inflated and spatial heterogeneity issues. Our findings suggest that streetscape features can effectively reflect built-environment characteristics at the road-segment level. Moreover, a comparison of our proposed modeling method with existing models demonstrates its superior performance. The results provide insight into the development of effective planning strategies to improve traffic safety.

Rapid land-use change detection (LUCD) is pivotal for refined urban planning and management. In this paper, we investigate LUCD through learning embeddings of points of interest (POIs) from multiple temporalities. There are several prominent challenges: (1) the co-occurrence problem of multi-temporal POIs, (2) the heterogeneity of POI categorization, and (3) The lack of human-crafted labels. Therefore, multi-temporal POIs need to be aligned in the embedding space for effective LUCD. This study proposes a multi-temporal POI embedding (MT-POI2Vec) technique for LUCD in a fully unsupervised manner. In MT-POI2Vec, we first utilize random walks in POI networks to capture their single-period co-occurrence patterns; then, we leverage manifold learning to capture (1) single-period categorical semantics of POIs to enforce semantically similar POI embedding to be close and (2) cross-period categorical semantics to align multi-temporal POI embedding in a unified embedding space. We conducted experiments in Shenzhen, China, which demonstrates that the proposed method is effective. Compared with several baseline models, MT-POI2Vec can better align multi-temporal POIs and thus achieve higher performance in LUCD. In addition, our model can effectively identify areas with unchanged land use and land use changes in residential and industrial areas at a fine scale.

A geospatial Knowledge Graph (KG) is a heterogeneous information network, capable of representing relationships between spatial entities in a machine-interpretable format, and has tremendous applications in logistics and social networks. Existing efforts to build a geospatial KG, have mainly used sparse spatial relationships, e.g., a district located inside a city, which provide only marginal benefits compared to a traditional database. In spite of the substantial advances in the tasks of link prediction and knowledge graph completion, identifying geospatial relationships remains challenging, particularly due to the fact that spatial entities are represented with single-point geometries, and textual attributes are frequently missing. In this study, we present GTMiner, a novel framework capable of jointly modeling Geospatial and Textual information to construct a knowledge graph, by mining three useful spatial relationships from a geospatial database, in an end-to-end fashion. The system is divided into three components: (1) a Candidate Selection module, to efficiently select a small number of candidate pairs; (2) a Relation Prediction component to predict spatial relationships between the entities; (3) a KG Refinement procedure, to improve both coverage and correctness of a geospatial knowledge graph. We carry out experiments on four cities' geospatial databases, from publicly-available sources and compare with existing algorithms for link prediction and geospatial data integration. Finally, we conduct an ablation study to motivate our design choices and an efficiency analysis to show that the time required by GTMiner for training and inference is comparable, or even shorter, than existing solutions.

Points-of-interests (POIs) have been proven to be indicative for sensing urban land use in numerous studies. However, recent progress mainly relies on spatial co-occurrence patterns among POI categories, which falls short in utilizing the rich semantic information embodied in POI hierarchical categories and in sensing the spatial distribution patterns of POIs at an individual zonal scale. In this context, we present a spatial and adversarial representation learning approach (SARL) for predicting land use of urban zones with POIs. SARL deeply mines the information from POIs from both spatial and categorical perspectives. Specifically, we first utilize a convolutional neural network to sense the spatial distribution patterns of POIs in each urban zone. We then leverage an autoencoder and an adversarial learning strategy to mine the POI categorical information in all hierarchical levels, which emphasizes the prominent and definitive POIs while preserves the overall POI hierarchical structures in each zone. Finally, we fuse these information from the two perspectives via a Wide & Deep network and carry out land use prediction with the fused embeddings. We conduct comprehensive experiments to validate the effectiveness of SARL in four European cities with real-world data. The results demonstrate that SARL substantially outperforms several competitive baselines.

The past decade has witnessed the increasingly created point-of-interest (POI) data, which are utilized to express the semantics of places. To understand the POI semantics, current methods usually embed POI categories into a latent space via certain trajectory sequential models, while neglecting the underlining spatial information. It is noteworthy that the complex spatial relationships among POI categories contain substantial information that benefits meaningful semantic embeddings for various POI categories. Inspired by this, we present a unified POI Category Embedding Method (CatEM for short), which jointly encodes the sequential transitions and spatial relations of POI categories as well as the adaptive semantic neighbors of each POI category. The merits of CatEM lie in that: (1) it considers the pairwise spatial similarities between categories and represents categories with larger similarity values with adjacent embeddings in the latent space, and (2) it adaptively locates neighbor categories with similar semantics in the embedding space to improve the adaptivity of POI category embedding. The proposed pre-trained POI category embeddings are justified by three downstream tasks. Extensive experiments demonstrate the superiority of our proposed model, as compared to several cutting-edge baselines.

We present a novel approach for estimating the proportional distributions of function types (i.e. functional distributions) in an urban area through learning semantics preserved embeddings of points-of-interest (POIs). Specifically, we represent POIs as low-dimensional vectors to capture (1) the spatial co-occurrence patterns of POIs and (2) the semantics conveyed by the POI hierarchical categories (i.e. categorical semantics). The proposed approach utilizes spatially explicit random walks in a POI network to learn spatial co-occurrence patterns, and a manifold learning algorithm to capture categorical semantics. The learned POI vector embeddings are then aggregated to generate regional embeddings with long short-term memory (LSTM) and attention mechanisms, to take account of the different levels of importance among the POIs in a region. Finally, a multilayer perceptron (MLP) maps regional embeddings to functional distributions. A case study in Xiamen Island, China implements and evaluates the proposed approach. The results indicate that our approach outperforms several competitive baseline models in all evaluation measures, and yields a relatively high consistency between the estimation and ground truth. In addition, a comprehensive error analysis unveils several intrinsic limitations of POI data for this task, e.g. ambiguous linkage between POIs and functions.

Next location prediction is of great importance for many location-based applications and provides essential intelligence to various businesses. In previous studies, a common approach to next location prediction is to learn the sequential transitions with massive historical trajectories based on conditional probability. Nevertheless, due to the time and space complexity, these methods (e.g., Markov models) only utilize the just passed locations to predict next locations, neglecting earlier passed locations in the trajectory. In this work, we seek to enhance the prediction performance by incorporating the travel time from all the passed locations in the query trajectory to each candidate next location. To this end, we propose a novel prediction method, namely the Travel Time Difference Model, which exploits the difference between the shortest travel time and the actual travel time to predict next locations. Moreover, we integrate the Travel Time Difference Model with a Sequential and Temporal Predictor to yield a joint model. The joint prediction model integrates local sequential transitions, temporal regularity, and global travel time information in the trajectory for the next location prediction problem. We have conducted extensive experiments on two real-world datasets: the vehicle passage record data and the taxi trajectory data. The experimental results demonstrate significant improvements in prediction accuracy over baseline methods.

Map mashups, as a common way of presenting geospatial information on the Web, are generally created by spatially overlaying thematic information on top of various base maps. This simple overlay approach often raises geometric deficiencies due to geometric uncertainties in the data. This issue is particularly apparent in a multi-scale context because the thematic data seldom have synchronised level of detail with the base map. In this study, we propose, develop, implement and evaluate a relative positioning approach based on shared geometries and relative coordinates to synchronise geometric representations for map mashups through several scales. To realise the relative positioning between datasets, we adopt a Linked Data–based technical framework in which the data are organised according to ontologies that are designed based on the GeoSPARQL vocabulary. A prototype system is developed to demonstrate the feasibility and usability of the relative positioning approach. The results show that the approach synchronises and integrates the geometries of thematic data and the base map effectively, and the thematic data are automatically tailored for multi-scale visualisation. The proposed framework can be used as a new way of modelling geospatial data on the Web, with merits in terms of both data visualisation and querying.