ABSTRACT Autonomous web-based geographical information systems (AWebGIS) aim to perform geospatial operations from natural language input, providing intuitive, intelligent, and hands-free interaction directly within a Web environment. However, most current solutions rely on cloud-based large language models (LLMs), which suffer from centralized processing issues like high operational costs and limited user privacy. This study explores an alternative architectural direction by deploying fine-tuned Small Language Models (SLMs) directly inside the user’s web browser to support AWebGIS. We compare two distinct approaches: (i) a method utilizing a cloud-based LLM (DeepSeek Chat V3.1), and (ii) a method based on three fine-tuned browser-executable SLMs – T5-small, T5-efficient-mini, and T5-efficient-tiny – executed entirely on the user’s device via WebAssembly/WebGPU. The T5-small model (60 M parameters, 155 MB) achieved an Exact Match Accuracy (EMA) of 0.82, Levenshtein Similarity (LS) of 0.96, ROUGE-1, and ROUGE-L scores of 0.95. Even the smallest model, T5-efficient-tiny (16 M parameters, 59 MB), maintained a functional EMA of 0.56. This browser-native, on-device strategy reduces the computational load on cloud servers, eliminates the need for large infrastructures to host LLMs, avoids token-based operational costs, and enhances user privacy. These results demonstrate the feasibility of leveraging highly compressed, browser-executable SLMs for developing cost-efficient, installation-free, and decentralized AWebGIS.

ABSTRACT Toponyms are crucial in providing spatial identities, as they characterize and give meaning and relevance to named geographical features. OpenStreetMap (OSM) has become a promising source of collaborative toponymic data, offering the potential to complement and update authoritative gazetteers by taking advantage of the dynamic nature of data updates and the local knowledge of contributors. However, updating official databases with collaborative data requires careful consideration and detailed understanding of the data quality. Although OSM is increasingly used in various geospatial applications, its role in supporting authoritative toponymic mapping remains relatively underexplored in the academic literature, particularly in terms of systematic methods for quality assessment. This study addresses this gap by presenting an open-source and reproducible framework capable of processing long-term OSM history data and extracting toponymic information across regular spatial grids, along with intrinsic quality indicators. The framework was applied in two different urban areas in Brazil, focusing on OSM features (amenity, building and leisure tags) with high potential for containing toponyms and semantic alignment with the Brazilian topographic mapping. The tool operates on 16 years of OSM data, demonstrating the spatial patterns of collaborative toponym contributions through the implementation of spatial statistical analysis. Moreover, the approach offers practical value for National Mapping Agencies (NMAs), providing a scalable method for assessing and potentially integrating collaborative toponyms into authoritative databases.

ABSTRACT Despite the implications of climate change, few Americans are engaged. Research suggests communicating through a more story-like structure increases engagement in climate change mitigation through “narrative transportation” – how engrossed one becomes in a story. In addition, map design research has recently touted the potential of storytelling maps. We report on two user studies examining the narrative structure, maps, and prior climate change beliefs on the dependent variables of: 1) narrative transportation, 2) attitudes, and 3) engagement. Using Geographic Information Systems (GIS) we explored the connection between these three variables and self-reported emotions. We found that participants who saw a more story-like narrative were more likely to be transported. When accounting for emotions, those who were less narratively transported reported feeling “fearful” or “hopeful,” while those who were more narratively transported were more “compassionate.” We found the inclusion of maps and photos did not make a difference, while prior beliefs were a significant predictor variable. These results add to the growing literature on the implications of how climate change information is conveyed, and confirms the importance of writing engaging stories.

ABSTRACT Choropleth maps are fundamental tools for geographic data analysis, primarily relying on color to convey information. Consequently, the design of their color schemes is important in choropleth map production. However, the traditional coloring methods offered by GIS tools such as ArcGIS and QGIS are not user-friendly enough for nonprofessionals. These tools provide numerous color schemes, making selection difficult, and cannot also easily fulfill personalized coloring needs, such as requests for “summer-like” map colors. To address these shortcomings, we develop a novel system that leverages a large language model and map color design principles to generate contextually relevant and user-aligned choropleth map color schemes. The system follows a three-stage process: Data processing, which provides an overview and classification of the data; Color Concept Design, where color theme and mode are conceptualized based on data characteristics and user intentions; and Color Scheme Design, where specific colors are assigned to classes. Our system incorporates an interactive interface for choropleth map color design and allows users to customize color choices flexibly. Through user studies and evaluations, the system demonstrates acceptable usability, accuracy, and flexibility, with users highlighting its efficiency and ease of use.

ABSTRACT When translating the names of geographical entities, it is often necessary to adopt specific translation strategies for geographical names that have a general name derivation relationship with other geographical names. Traditional rule-based methods are insufficient for identifying widespread generic name derivation relationships between geographical entities. This paper proposes a method for identifying general name derivation relationships based on prompt learning and Semantic-Spatial Distribution Attention Feature Fusion (SSD-AFF) to address this issue. This method aims to improve the efficiency and accuracy of identifying general name derivation relationships between place names. This method is divided into two steps: 1) Identification of derived place names. We use prompt learning methods, where the generic name derivation pattern serves as the context for place names, allowing a bidirectional language model to classify the place names. 2) Identification of derivation relationships. We utilize a Semantic-Spatial Distribution Feature Fusion module (SSD-AFF) to integrate various features of generic name derivation relationships, thereby achieving the identification of such relationships between place names. Experiments show that this method outperforms traditional methods in terms of performance and robustness and has certain application value in bilingual map production.

ABSTRACT Historical maps represent an invaluable memory which should be preserved. Such kind of maps are, however, mostly scanned and stored as raster graphics which do not contain semantic information in a machine-readable form. To achieve a machine-readable state, an often expensive human intervention is needed in a fully manual or semi-automatic fashion. An automatic interpretation and a feature extraction is then inevitable for a map digitization and vectorization. Automatic approaches showed more and more convincing and promising results on challenging map corpora avoiding human interaction. This paper deals with the semantic segmentation of historical maps based on Graph Convolutional Networks (GCNs) to capture long-range dependencies between image features. This allows for an extension of the receptive field of Convolutional Neural Networks (CNNs) restricted on local dependencies. A Self-Constructing Graph (SCG) module has been applied to automatically induce the structure of the GCN. We performed experiments revealing promising results where our approach achieved an Mean Intersection over Union (mIoU) of 0.68 , outperforming a state-of-the-art CNN dedicated to the semantic segmentation of historical maps.

ABSTRACT Selection of urban road networks has always been the focus of map generalization. Previous research has shown that machine learning techniques effectively address multi-attribute weighting challenges. However, these studies primarily focus on supervised learning that requires large amounts of labeled data and scale-specific training, limiting the automation of road network selection processes. To address these challenges, this study proposes an automated selection of urban road network. The method improves the automation level of the selection process by fusing the PageRank algorithm and attribute importance metrics and considering both the structural and attribute information of roads. The proposed method was tested on four typical road network cases from China and compared with the single-attribute decision method, multi-attribute decision method and expert-selected results. The results show that this method can preserve the connectivity and density distribution of the source map.

ABSTRACT The reproducibility and replicability (R&R) crisis poses a significant challenge across disciplines, particularly in spatiotemporal studies. This paper focuses on the unique challenges within spatiotemporal research in the context of R&R, including data availability, methodological conception transparency, interdisciplinary collaboration complexities, the balance between R&R and innovation, and R&R education. Recognizing the potential of Scientific Workflow Management Systems (SWMS) to enhance R&R, we introduce a pioneering SWMS-based integrated spatiotemporal research approach (SISRA) utilizing KNIME, an open-source SWMS, to tackle these R&R challenges. First, we developed a set of KNIME extensions, including Geospatial and Dataverse extensions, to enhance spatiotemporal software availability in SWMS. Then we created spatial data virtual laboratory architecture to support multidisciplinary collaboration. Finally, we suggested a geographical research lifecycle that integrates SWMS-based methods to improve practices, efficiency, and innovation in R&R research and education. Our approach exemplifies how executable workflows can not only alleviate the R&R burden on researchers but also strengthen R&R education in geographical research, illustrating the benefits of our approach in training, teaching, and multidisciplinary collaboration.

ABSTRACT This paper presents a new methodology for identifying copying relationships in maps through cartographic deformation analysis, and introduces a novel approach for the creation of synthetic datasets for evaluation. Challenging traditional approaches that compare historical maps to modern ones for an accuracy assessment, this study focuses on the comparison between old maps through a modern reference map to identify copying processes. Using a synthetic dataset of deformed maps, this research introduces a novel way to assess deformation analysis results. The methodology’s effectiveness is demonstrated through experiments on the synthetic dataset and on five maps of Jerusalem where copying relationships are known, revealing that cosine similarity, filtered with vectors’ norm and combined with their ratio, offers a good discriminative power between maps that are complete copies and those which are not. The methodology, with the selected metric, is applied to a selection of fifty 19th-century maps of Jerusalem and used to create a hierarchical clustering. This approach enables the identification of complete copies of varying quality within large map collections. This study not only advances cartometric methodologies but also lays the groundwork for future research on the automatic detection of map lineages, marking progress toward the genealogical analysis of large cartographic datasets.

ABSTRACT Spatial knowledge, especially survey knowledge, plays a pivotal role in people’s daily spatial tasks. However, the inherently limited visual perception and integration of spatial information often leads to inaccuracies in survey knowledge, with far-reaching consequences that affect performance, emotion, and safety in completing spatial tasks. In this study, we developed a mixed reality (MR)-based coordinate system and conducted a comparative user experiment involving 56 participants to explore whether perceiving metric and global spatial references during navigation may enhance the accuracy of user-acquired spatial knowledge. The experiment encompassed a spatial learning task where participants navigated and learned the selected indoor environment, followed by a desktop-based landmark recalling task to evaluate the accuracy of participants’ acquired survey knowledge. The results revealed the following insights: (1) the use of MR-based coordinate system significantly improved egocentric and allocentric spatial relations in participants’ survey knowledge; (2) participants’ errors in spatial relations among landmarks accumulate over distance, whereas the MR-based coordinate system helps maintain the errors at a consistent level; (3) no significant correlations were found between errors in participants’ survey knowledge and environmental familiarity, suggesting that the aiding effect of the MR-based coordinate system was independent of users’ prior knowledge of the environment; (4) participants reported increased mental demand using the MR-based coordinate system, but this did not result in diminished spatial learning performance or additional time pressure and learning efforts. These empirical results contribute to the future study of spatial knowledge acquisition and the design of spatial learning interfaces in mixed reality.