iiinova — where your digital self keeps evolving

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 17/09/2026

CaGIS Volume 53 Issue 5, including a special section on AsiaCarto 2024 is now available online and the printed copy is on its way to our subscribers! www.tandfonline.com/toc/tcag20/5... #GISchat We have 8 great papers, check out the thread below:

Front cover, for Cartography and Geographic Information Science, Volume 53, No 5, September 2026, The Journal of the Cartography and Geographic Information Society. Includes Figure 1 by Yang et al. A sample of map layout graph construction, which is a research diagram illustrating map layout graph construction. Part a) shows a world map of seismic activity with 6 colored markers representing different elements on the map. Part b) transforms these into a graph theory representation where elements become vertices (v1-v6) connected by edges showing relationships.

122

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 03/08/2026

New article! Fascinating article from Lu Cheng, Taisheng Chen and colleagues exploring differences of spatial visual attention and spatial confidence during spatial orientation in unfamiliar environments doi.org/10.1080/1523... #GISchat

Figure 3: Representative composite heat maps of visual attention during map-based navigation tasks 2-4, comparing a) a poor-SD (spatial disorientation) participant and b) a good-SD participant. Both maps show the same urban area with streets, parks (green), water (blue), and numbered points of interest (1-10 and A-D). Starting and end points are labeled. Heat map colors indicate visual attention intensity: orange/red shows high concentration, blue shows low or no attention. The poor-SD participant's map displays concentrated orange/red clusters throughout, suggesting scattered attention. The good-SD participant's map shows dispersed blue tones with subtle patterns, suggesting more systematic scanning. Each map includes a north arrow and 1:20m scale bar. The comparison demonstrates differences in visual attention patterns related to spatial disorientation abilities during map-based wayfinding.

012

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 22/07/2026

Check out the CaGIS Interview with Mahdi Nazari Ashani to hear all about running a SLM in a browser youtu.be/YZ-BFNHtwMo and checkout the full demo of AWebGIS at youtu.be/7-QYALy4F2E plus the CaGIS article for #free at doi.org/10.1080/1523... #GISchat

youtu.be

CaGIS Interview: Mahdi Nazari Ashani on Small Language Models for AWebGIS

YouTube video by Cartography and Geographic Information Society

030

Reposted by The Cartography and Geographic Information Society

Paolo Beria 🇪🇺 @beriapaolo.bsky.social · 15/07/2026

For example we point out places with "monodirectional" mobility - such as metropolitan fringes dependent from highly attractive cores - or areas whose mobility is predominantly erratic. We applied the method to Italy and England+Wales and tested also a diachronic version (delta between two years)

011

Reposted by The Cartography and Geographic Information Society

Paolo Beria 🇪🇺 @beriapaolo.bsky.social · 15/07/2026

Out our new paper on @carto-giscience.bsky.social, full of nice mobility maps! More seriously: we created a novel (and relatively easy) way to map entire OD matrices, describing general mobility patterns of each area.

111

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 14/07/2026

New article! Federico De Angelis & Paolo Beria @beriapaolo.bsky.social present some new ways of mapping origin-destination matrices with national-scale commuting data doi.org/10.1080/1523... #GISchat

Mobility resultants for Italy, work commuting, year 2011. Labels refer to places mentioned in the text. The map displays Italy with terrain shown in light gray and major cities labeled including Milan, Turin, Venice, Bologna, Rome, Naples, Palermo, and others. Orange flow lines represent commuting patterns between locations, with line thickness indicating volume of movement. The legend shows two classification systems: 'Monodirectionality index' (0-0.33, 0.33-0.50, 0.50-0.66, 0.66-1.00) and 'Total movements [pax]' (1-257, 257-614, 614-1577, >1577). Strong commuting corridors appear around major urban centers like Milan and Rome, shown by prominent orange flow patterns radiating outward. Regional annotations identify areas like Rimini, Ancona, Cagliari, and Bari. The visualization demonstrates uneven geographic distribution of work commuting, with concentrated patterns in northern industrial regions and more dispersed patterns in southern Italy.

074

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 30/06/2026

CaGIS Volume 53 Issue 4 is now available online and the printed copy is on its way to our subscribers! www.tandfonline.com/toc/tcag20/5... #GISchat We have 8 great papers looking at visualisation, spatial modelling and AI in GIS. Check out the thread below:

Front cover, for Cartography and Geographic Information Science, Volume 53, No 4, July 2026, The Journal of the Cartography and Geographic Information Society. Includes Figure 6 by Yang et al.: (in this issue) System interface (an example of “Statue of Liberty” map color design). a) Conversation view: users can input their initial design intents and customizations in natural language. b) Color design view: users can fine-tune the intermediate results generated by the LLM. c) Map view: users can evaluate the results with reasoning and refine the color assignments based on their preferences.

130

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 29/06/2026

New article from Kurtis Danluk, Bernhard Jenny and colleagues presenting Ring maps: a new view for augmented reality navigation doi.org/10.1080/1523... #GISchat Source code at github.com/berniejenny/...

Figure 1: A ring map with an inner and an outer ring created by the prototype for the Apple Vision Pro augmented-reality headset, as seen from the viewer's perspective. The photograph shows an outdoor plaza area with trees, buildings, and people in winter/early spring conditions. A hand holding the device is visible in the foreground pointing toward an augmented reality overlay. The AR visualization displays concentric rings labeled 'Inner ring' (Closest landmark distance 10 m) and 'Outer ring' (Farthest landmark distance 500 m). A 3D landmark model of McKinney Tower is overlaid on the real-world landmark building in the center distance. White text labels identify the real-world landmark, the 3D landmark representation, and the ring boundaries. A pink/magenta colored rectangular shape appears on a nearby tree trunk, labeled 'Location & Heading +15'. The visualization demonstrates spatial awareness and proximity measurement in augmented reality.

022

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 25/06/2026

New article! Atsushi Masuyama explores execution-level variability for geographic masking: a new way to help improve privacy without compromising analytical utility doi.org/10.1080/1523... #GISchat Check out the data at doi.org/10.6084/m9.f...

Figure 3: Comparison of crime locations before and after masking using donut masking with different parameter settings. Panel a) shows results with r = 100 m and panel b) with r = 500 m. Both maps display the same urban area with underlying street network visible as thin gray lines, census blocks (solid gray polygons labeled 'Households ≥ 0'), and Chôme boundaries (white/unfilled polygons). Pre-masking Hypothetical Crime Locations are marked with filled black circles, representing original crime positions. Post-masking Crime Locations are shown as hollow/open circles, indicating the displaced positions after applying the donut masking algorithm. Each map includes a north arrow and 1 km scale bar. The larger radius in panel b) (500m) produces greater displacement of crime locations compared to the smaller radius in panel a) (100m), demonstrating how the radius parameter controls the privacy protection level in crime data masking.

021

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 22/06/2026

New article! Shreya Shukla & Tina Pujara complete a systematic review of how we cartographically represent emotions in urban design and planning doi.org/10.1080/1523... #GISchat

Figure 3: Link between the methods of emotion data collection and visualization techniques. The diagram displays three columns connected by arrows showing data flow. Left column 'Methods of Emotion Data Collection' is divided into three categories: Subjective (Map-based survey, Mental map/Sketch Map, Questionnaire), Crowdsourcing Subjective (Crowdsourcing Social media data), and Objective (Physiological sensors). Center column 'Data Type' shows intermediate processing stages: Data lines/segments, Sketches/Annotations, Geolocated Data points (with note 'Combined with GPS data'), and Continuous data values. Right column 'Visualization Techniques in Emotion Mapping' displays output maps: Line Map, Sketch Map, Emoji Map, Point Map, Pie-chart Map, and Peak Map. Additional processing step shows 'Data Aggregation' connecting Point Map and Pie-chart Map to Heat Map and Grid Map respectively. Dashed and solid arrows indicate direct and conditional connections between collection methods, data types, and visualization outputs.

011

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 18/06/2026

New article! Christos Kastrisios proposed a new side selective line simplification algorithm for use with nautical chart generalization doi.org/10.1080/1523... #GISchat #OpenAccess See the data at doi.org/10.6084/m9.f...

Figure 14: Representative sections from testbeds a) 6 (left), b) 4 (center), and c) 2 (right) showing removal of source-line intricacies and preservation of overall contour form by the side-selective algorithm. Each panel displays red contour lines on a white background, representing different results of line simplification processing. Panel a) shows contours with parallel linear features and some detailed intricacies. Panel b) displays a more centralized pattern with less peripheral detail. Panel c) exhibits scattered, dispersed contour segments. The three examples demonstrate varying degrees of how the algorithm handles the removal of fine-scale variations and noise from source contour data while maintaining the essential shape and spatial organization of the overall landform representation, making note of which side can be simplified.

021

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 17/06/2026

youtu.be

CaGIS Interview: Mahdi Nazari Ashani on Small Language Models for AWebGIS

YouTube video by Cartography and Geographic Information Society

000

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 09/06/2026

021

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 04/06/2026

New article! Fascinating article from Rui Xin and colleagues, looking at using AIS data to identify occupied anchor positions in ports, using a case study of Port of Los Angeles - Long Beach #GISchat doi.org/10.1080/1523... See their code at doi.org/10.6084/m9.f...

Variation of the average offset (movement due to weather and tides) at various time points for Long Beach. Darker red = more offset.
Figure 18: Visualization of the average offset degree of occupied anchor positions from 2019 to 2023. Six maps of the Long Beach area arranged in a 2x3 grid show: a) combined five-year data (2019-2023), and b) 2019, c) 2020, d) 2021, e) 2022, f) 2023 individually. Anchor positions are displayed as circles colored on a heat scale from light pink/white (0, no offset) to dark red (1, maximum offset), indicating the average displacement of occupied anchors. Spatial clustering patterns are visible, with higher offset degrees (darker red) in southern zones during earlier years, decreasing toward 2023. Each map includes a north arrow and scale bar (0-2-4 km). Light blue/teal represents water and beige represents land.

031

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 20/05/2026

New article! Xini Hu and colleagues look at how we can create a map color style design to enhance the emotional expressiveness of maps #GISchat doi.org/10.1080/1523...

Figure 7: Sample cases of emotional color style design results for ordinary maps. Six panels labeled a) through f) display partial Chengdu city vector maps styled with different emotional color palettes. Each panel shows a small reference image (left) with an arrow pointing to the resulting styled map (right). Panel a) 'original' shows the standard map with typical cartographic colors. Panel b) 'calm' uses muted beiges and soft colors. Panel c) 'happy' features bright teals, greens, and warm tones. Panel d) 'nostalgic' employs sepia and brown tones. Panel e) 'fear' uses dark blues and purples creating an ominous atmosphere. Panel f) 'depressed' applies desaturated grays and muted colors. Each map includes a legend identifying features: river, forest, residential, industrial, commercial, main road, subsidiary road, ordinary road, railway, and background. The maps show the same geographic area with consistent road networks and blue river running through, demonstrating how color choices alone can evoke different emotional responses.

073

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 18/05/2026

New article! How can we calculate off-road trafficability? Renan Fabres Dalmonech and colleagues use multicriteria decision analysis to develop a robust spatial model #GISchat doi.org/10.1080/1523...

Figure 6: Vehicle trafficability map for a conventional vehicle with tires, showing a) a continuous scale ranging from 0 to 10 and b) a categorical scale with classes from adequate to impeditive, according to Jenks natural breaks. Both maps display the same geographic area with coordinate labels (27.92°S to 28.06°S latitude, 49.44°W to 49.36°W longitude). Map a) uses a continuous color gradient from light orange/beige (low trafficability, 0) to dark brown/red (high trafficability, 10) as shown in the Index of Trafficability (IT) legend. Map b) applies five distinct color categories: green (Adequate), light green (Slightly restrictive), yellow (Restrictive), orange (Very restrictive), and red (Impeditive). The right panel includes a legend with both classification systems, a north arrow, scale bar showing 0-5-10 km at 1:10,000 scale, and datum information (Horizontal Datum SIRGAS2000 Epoch 2000.4, UTM Zone 22S). The terrain appears highly variable with complex patterns of trafficability throughout.

011

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 13/05/2026

CaGIS 53-2 special issue on Replicability and Reproducibility is now available online! www.tandfonline.com/toc/tcag20/5... #GISchat We have 5 great papers looking at reproducibility, including in data acquisition, education, map reproduction and research process. Check out the thread below:

Front cover, for Cartography and Geographic Information Science, Volume 53, No 3, May 2026, The Journal of the Cartography and Geographic Information Society. Includes composite figure from Holler et al. (this issue) including Figure 3 (center) Network graph of studies and versions, referencing maps from
different studies including the original study material, Figure 2 (Kang
et al, top-left), the Chicago Reproduction, Figure 4 (top-middle) and
Reanalyses of spatial access in Class Projects, Figure 7 b) - e)
(top-right and bottom).

121

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 11/05/2026

New article! Fascinating article from Laura Wenclik and Guillaume Touya: Staring at the pointer - gaze behavior when people zoom and pan in maps doi.org/10.1080/1523... #GISchat Check out their data at doi.org/10.5281/zeno...

Figure 5: Example of a linear zoom, where the gaze follows a linear translation regarding the map location during the zoom. The image shows a map background (from OpenStreetMap) with place names like Nevillers, Fouqueroles, and Le Fay-Saint-Quentin visible. Multiple translucent green circles of varying sizes are positioned along a roughly horizontal black line crossing the map from left to right. Each circle represents a fixation point with a radius related to the zoom level of the map at the time of the fixation - larger circles indicate higher zoom levels (zoomed out view), while smaller circles indicate lower zoom levels (zoomed in view). The different shades of green represent different interactions that occurred during the map exploration. A scale bar showing '1500 m' appears in the lower left corner. The circles overlap in sequence, creating a visual trail of the user's gaze path and zoom behavior during map navigation.

031

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 07/05/2026

Fantastic article by Carolyn Fish, reflecting on how we contextualise cartography in a changing climate and her journey from enthusiasm to disillusionment to a reignited excitement for maps, from her CaGIS Early Career Scholar keynote at AAG CMSG in Mar 2025. doi.org/10.1080/1523... #GISchat

The changing climate of internet cables: Figure 7: Section of map titled 'Deep-Sea Internet Cables and the Ownership of Global Infrastructure' by University of Oregon student Michal Klopotowski, illustrating the increasing control of undersea cable by Amazon, Google, Meta, and Microsoft. The map displays North and South America with terrain relief, surrounded by dark blue oceans. Submarine cable routes are shown as curved lines radiating from coastal landing points (marked with yellow dots) across the Pacific and Atlantic oceans. Cables owned by content providers (Amazon, Google, Meta, Microsoft) are distinguished by color: orange/yellow for existing cables and dashed orange/yellow for planned cables (2025+). Other telecom provider cables are shown in white/light blue for existing and dashed white/light blue for planned routes. An ocean depth gradient scale ranges from 10,000m (dark blue/black) to 0.0m (light blue/white). Credit: Cartographer Michal Klopotowski.

0135

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 06/05/2026

New article! Christos Kastrisios and Colin Ware propose a new way of incorporating uncertainty into bathymetric data visualisations doi.org/10.1080/1523... #GISchat #OpenAccess Check out their data at doi.org/10.6084/m9.f...

Figure 5: SimChart example showing a plotted route, color coding of chart features, and tested mode overlay (here, the Hazard mode) used in the experiment. The image displays an artificial nautical chart with light blue water areas marked with depth soundings (numerical values like 10₉, 11₂, 16₄). Shallow depth areas labeled 'Shallow Depth Area' are shown in tan/beige color. A yellow line labeled 'Safety Contour' runs horizontally through the middle section. Red arrows indicate a plotted navigation route. The chart includes diagonal hatching pattern in the background and various depth contour lines. In the upper right corner, a legend shows 'HJ DU' (Hazard) depth categories: 0-5m (white), 1-5m (light blue), 1-5m (darker blue), and 3-0m (darkest blue), with a scale numbered 1 through 5. Brown/gray areas appear to represent land or very shallow regions. The window title shows 'Artificial Chart' with standard minimize, maximize, and close buttons.

021

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 30/04/2026

New article! Ondřej Kvarda and colleagues test the impact of different visualisation methods - bar graphs (extrinsic) and Chernoff faces (intrinsic) - and map literacy - on user performance within immersive virtual reality (IVR) doi.org/10.1080/1523... #GISchat #OpenAccess

- Figure 3 showing two panels comparing different visualization approaches for the same spatial analysis task. Both panels display a 3D isometric view of a raised platform with colored rectangular blocks representing data values, along with interface elements for task instructions and legend information.
- Top panel: Shows Task 14/15 labeled 'Úkol 14/15' with instructions in Czech ('Identifikujte území s Žlutou hodnotou Proměnné1 a zároveň Vysokou hodnotou Proměnné2') and English ('Identify an area with a high value of Variable1 and a high value of Variable2'). The legend is organized in a traditional tabular format showing: Variable1 with categories Proměnná1 (items 1-4) in shades of blue, and Variable2 with categories Proměnná2 and 'nízká střední vysoká' (Low, Medium, High) in shades of red. The 3D platform shows four numbered areas (1-4) with colored blocks representing different combinations of the two variables. An 'Answers' section shows 'Odpovědi' with numbered options 1-4.
- Bottom panel: Shows the same Task 14/15 with identical instructions, but presents an alternative legend design. Instead of a traditional table, the legend uses a more integrated spatial layout with Variable1 categories (Proměnná1) and Variable2 categories (Low, Medium, High) arranged in an overlapping format. Additional emoji-like face icons appear in the upper right area showing various expressions (neutral, sad, happy faces) arranged in a grid. The 3D platform shows the same four numbered areas (1-4), but now includes small face icons positioned above certain blocks. The 'Answers' section (Odpovědi) again shows numbered options 1-4, with Variable2 category labels (Low, Medium, High) and Proměnná2 visible.

021

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 28/04/2026

New article! Mengjun Kang, Tianyi Bai and colleagues explore matching geographic names (toponyms)from both textual and spatial features doi.org/10.1080/1523... #GISchat

Diagram titled “Learning process for the representation of toponyms,” with three sections: Linguistic symbol system (left), High-dimensional vector space (center), and Application (right). In the left section, a box with “name,” “category,” and “…” feeds into “Natural language text,” which is converted via word embedding into a “Word vector.” A “Toponym” (text feature) connects via dashed “Vector mapping” to the center. Below, a coordinate system with “location,” “pattern,” and “…” feeds into “Euclid geometry,” then through location encoding into a “Location vector.” In the center, “Word vector” and “Location vector” combine into a “Multimodal joint representation,” producing a “Toponymic vector.” In the right section, this vector supports applications: data fusion, toponymic matching, and knowledge graph.

031

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 01/04/2026

New article! Fantastic article from Mahdi Nazari Ashani and colleagues, showing how a SLM (Small Language Model) running a GIS can be run solely in a web browser tab with no data shared externally doi.org/10.1080/1523... #GISchat See their code at github.com/mahdin75/awe...

Figure 6: Example of autonomous GIS operation where the fine-tuned SLM translates natural language queries into executable functions (addWFS and SetStrokeColor), updating the map visualization in real-time. The screenshot shows the AWebGIS App interface with three main sections: 1) A left sidebar containing 'Analysis Tools' and 'Layer Manager' panels, with options to 'Add WMS', 'Add WFS', 'Import', and 'Add Base', plus layer options for 'states', 'wfs', 'OpenStreetMap', and 'base'; 2) A central map display showing a base map of North America with the United States outlined in red stroke, including state boundaries and the Great Lakes region; 3) A right sidebar with 'Chat Assistant' and 'Activity Log' tabs, showing a conversation thread with timestamps. The chat displays natural language commands such as 'Can you add 'states' layer from [http://localhost:8080/geoserver/cite/wfs](http://localhost:8080/geoserver/cite/wfs)?' (08:19 PM), followed by 'Executed: addWFS('states', '[http://localhost:8080/geoserver/cite/wfs](http://localhost:8080/geoserver/cite/wfs)', 'states')' (08:18 PM), then 'Set the color of the strokes to 'red' for 'states'' (08:19 PM), and 'Executed: SetStrokeColor('states', 'red')' (08:18 PM). At the bottom, the Active Model shows 'T5 Tiny-fine-tuned, Local' with a prompt to 'Ask me to perform GIS operations...' The interface includes top navigation buttons for 'GitHub', 'Print Map', and 'Full Screen'.

Figure 2: Conceptual workflow of the autonomous web-based geographical information systems (AWebGIS) approaches. The diagram shows a natural language query at the top splitting into two parallel processing paths. Approach I (Cloud-based LLMs), shown on the left in a dashed box, follows this sequence: 1) Natural-language query sent via HTTP to cloud API (OpenRouter to DeepSeek V3.1); 2) In the cloud, DeepSeek Chat V3.1 translates the query into a GIS function call using few-shot learning; 3) Predicted function call is returned as a text string to the web browser; 4) The JavaScript engine parses the function call and executes the corresponding GIS operation. Approach II (Browser-executable SLMs), shown on the right in a dashed box, follows this sequence: 1) Natural-language query processed locally through fine-tuned SLMs (e.g., T5-tiny); 2) In the user's web browser, an SLM translates the query into a GIS function call; 3) The JavaScript engine parses the function call and executes the corresponding GIS operation. Both approaches ultimately result in JavaScript execution of GIS operations, but differ in where the natural language processing occurs (cloud versus local browser).

042

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 31/03/2026

New article! Tianyang Bai, Weihua Dong and colleagues investigate Age-related deficits in reference frame switching of navigation ability, using fMRI doi.org/10.1080/1523... #GISchat

Figure 2: Results of the functional connectivity analysis of different age groups. Panel a) shows a circular chord diagram representing the functional connectivity matrix during Task 1, with brain regions labeled around the perimeter (PCC, PCA, MFG, UifG, MOG, PCUN, dlSFG, dSFG, PCUN, MFG) and small brain images positioned outside the circle. Connections between regions are shown as ribbons colored in pink (elderly greater than young, p < 0.01 and p < 0.005) and purple (elderly less than young, p < 0.01 and p < 0.005). Panel b) displays a similar chord diagram for Task 2, showing connections between regions including dlSFG_L, dlSFG_R, LinG, PCC, PCC, and MFG_R. Panel c) contains three box plots comparing elderly (pink) versus young (purple) groups: shortest path length for Task 1 and Task 2 (marked with ** and * for significance), normalized path length for Task 1 and Task 2 (marked with *), and global efficiency for Task 1 and Task 2 (marked with ** and *). Panel d) shows nodal metrics of network analysis, featuring two brain renderings (left and right hemispheres) with highlighted regions in purple (dlSFG_L and dlSFG_R for dorsolateral superior frontal gyrus). Four box plots surround the brain images comparing elderly versus young groups for nodal clustering coefficient and nodal local efficiency in both left and right dlSFG regions. Legend indicates: PCC=posterior cingulate cortex; LinG=lingual gyrus; MFG=middle frontal gyrus; MOG=middle occipital gyrus; PCUN=precuneus.

021

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 30/03/2026

New article! Zeyad Kelani and colleagues collate Point of Interest data (POI) from a variety of open sources for use in Drug and Substance Abuse work and find that these are comparable with commercial POI data sets doi.org/10.1080/1523... #GISchat

031

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 26/03/2026

New article! Kejia Huang, Niaz Muhammad and colleagues propose Geospatial Simultaneous Localization and Mapping for Outdoor Mobile Augmented Reality (GSOMAR), enabling real-time modeling and interaction in complex urban environments doi.org/10.1080/1523... #GISchat

Figure 13: Based on the proposed method, operators can dynamically view matching results from different positions and perspectives. Panel a) shows a long-distance perspective with a person in military-style clothing holding a tablet device outdoors on dusty terrain, with another person visible in the background. Panel b) displays a close-range perspective with hands holding a tablet under a concrete bridge structure, showing the device screen displaying augmented reality content aligned with the physical surroundings. Panel c) presents actual screenshots of geo-matching results featuring a 3D yellow wireframe model of bridge infrastructure overlaid on a photograph of gray concrete bridge support columns and beams, with text reading 'After geo-matching of bridge' at the bottom of the overlay.

Figure 15: Interactive spatial perception in outdoor mobile augmented reality (MAR). At approximately 40 meters from the building, users assess the alignment between virtual and real structures and query semantic attributes. Panel a) shows accurate alignment in position and orientation, displaying an aerial/top-down view of a white building with blue-tinted windows and a 3D model overlay. Panel b) shows GIS slicing revealing full indoor semantics, with a ground-level view of the building facade overlaid with numerous colored geometric icons (purple diamonds, orange hexagons, green squares, blue circles) representing different semantic attributes throughout all floors. Panel c) displays fourth-floor attributes including offices, restaurant, and restrooms, showing the same building view with fewer colored icons concentrated on specific floors. Panel d) presents first-floor attributes including meeting rooms, restrooms, and mother-and-baby room, with blue and green icons visible primarily on the lower portion of the building. Yellow flowers are visible in the foreground of panels b), c), and d), with palm trees and additional buildings visible in the background.

032

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 24/03/2026

New paper! Colorful map draining your phone battery? Hongyue Zhang and colleagues evaluate how to balance user preference, visual quality and energy efficiency in maps for mobile devices doi.org/10.1080/1523... #GISchat

In Figure 13 below: a) Best energy efficiency; b) best user preference; c) best visual quality; d) e) closest to ideal point; f) optimal solution.
Figure 13 showing six different cartographic styling variations of the same urban area map. Each panel a) through f) displays an identical street network with yellow/orange/brown road lines overlaying land parcels in various colors including green spaces (parks), blue water bodies, and urban blocks in pink, purple, cyan, and other hues. Panel a) features a dark teal/navy background with bright contrasting colors. Panel b) shows an olive/brown background with softer tones. Panel c) has a light beige/cream background with pastel colors. Panels d), e), and f) all use black backgrounds with progressively different color palettes: d) uses yellow streets with bright multi-colored parcels, e) employs pink/magenta streets with rainbow-hued parcels, and f) features green streets with a cyan-green-red color scheme. The maps demonstrate different approaches to urban cartographic design and color theory while maintaining the same geographic information.

052

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 18/03/2026

New article! Yassmine Zada, Eric Guilbert and Sylvain Jutras explore how to integrate culverts into drainage network mapping without altering the underlying digital terrain model (DTM) doi.org/10.1080/1523... #GISchat

Figure 10: Dale of a surface watercourse, with note that no dale is computed for the culvert. The map displays a grayscale elevation raster with drainage network overlay. Red lines indicate existing drainage divides forming a complex network across the terrain. Orange lines show a proposed new drainage divide boundary. Blue lines represent existing surface watercourses, while a thick cyan line shows a new proposed surface watercourse. A thick green line indicates a culvert location. Critical points are marked with various symbols: inverted blue triangles for pits (local elevation minima), cyan diamonds for confluences (stream junctions), yellow circles for saddles (low points along ridges), green circles for transfluences (water flow across divides), and red triangles for pics (local elevation maxima). A scale bar shows distances in kilometers from 0.01 to 0.03. The legend on the right defines all symbols and line types.

042

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 16/03/2026

CaGIS Volume 52 Issue 2, our special issue on Smart cartography for sustainable development: International Cartographic Conference 2023 is now available online and the printed copy should be on its way to all our subscribers! www.tandfonline.com/toc/tcag20/5... #GISchat

Front cover, for Cartography and Geographic Information Science, Volume 53, No 2, March 2026, The Journal of the Cartography and Geographic Information Society. Includes Figure 6 from Gołębiowska et al. Tested stimuli: a) 2D, b) 3D, c) and context in continuous (top row) and discrete (bottom row) RC. The images show visualizations of the subsurface temperature field and permafrost boundary inside the Zugspitze mountain in Germany. Data source: Noetzli et al. (2010), Gallemann et al. (2017), Deutscher Wetterdienst DWD, Geobasisdaten © Bayerische Vermessungsverwaltung 2011

152

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 12/03/2026

New article! Fantastic work from Nihal Z. Miaji and colleagues on the cartogram creation process, and how we ensure topology is preserved and cartogram regions remain connected and not overlapping doi.org/10.1080/1523... #GISchat #OpenAccess

Figure 2. Illustration of line densification when approximating a curve with a finite number of points. Panel (a) depicts the ground
truth, showing the precise geometry of two curves, C1 and C2. Panel (b) illustrates finite approximations, A1 and A2, that use a sparse
distribution of points along the respective curves. Topology is violated as A1 intersects itself and also A2. Panel (c) demonstrates line
densification resulting in A0 1 and A0 2 preserving the true topology

072

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 11/03/2026

New article! Wenwu Tang and colleagues present A web-based spatial decision support system of COVID-19 wastewater surveillance on a university campus doi.org/10.1080/1523... #GISchat

Figure 7. Snapshot of the Web GIS dashboard (sewer networks are hidden due to confidentiality considerations. Screenshot of a Web GIS dashboard showing sampling results over a campus map. The central panel displays a map with colored building footprints and sampling locations marked with colored dots indicating results (negative, positive, suspicious, or other). A legend on the left explains the symbols for samplers and buildings. A sidebar allows neighborhood selection and zooming. On the right, panels show selectable basemaps, a summary indicator reporting **25 positive individual sites** in the current map extent, and a pie chart showing the composition of testing results (not collected, negative, positive, suspicious). The bottom section contains a bar chart showing the number of positive samplers over time and a line chart showing the time-series testing results for a selected sampler.

071

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 10/03/2026

New article! Hua Liao and colleagues evaluate how much we can deduce from users visual behaviour in pedestrian navigation: quite a bit it seems, inc. gender, geographical expertise, spatial ability & familiarity of the environment #GISchat doi.org/10.1080/1523... Data at figshare.com/s/5b06a12bea...

Figure 1: User attributes and navigation behavior. The diagram shows a central pedestrian figure with four attribute categories branching outward: 1) Gender - showing female (illustrated with person receiving directions 'Turn left before the red building') and male (illustrated with person at directional signpost receiving instruction 'Go west for 300 meters'); 2) Expertise - divided into Geography (showing topographic contour map with task 'Find the highest point in the terrain') and Non-geography (showing street map with task 'Find the hospital on Queen street'); 3) Spatial ability - showing high spatial ability (person solving maze saying 'Easy!') and low spatial ability (confused people asking 'Where is...?'); 4) Familiarity - showing familiar environment (person in park saying 'I have been there!') and unfamiliar environment (person consulting map with 'Map told me!'). Below these attributes are five navigation task types illustrated with icons: Self-localization (map with location pin asking 'Where am I?'), Object search (red house with 'Find the house with red roof'), Map target search (map showing 'Find the picnic point on the map'), Route memorization (map with dotted route to Bill's Wood), and Walking to the end (person walking in urban environment).

062

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 09/03/2026

New article! Ben Ma and colleagues evaluate how mid-air gestures could be used to interact with mobile maps, doi.org/10.1080/1523... #GISchat Data available at doi.org/10.6084/m9.f...

Table 2: Mapping between tasks and candidate gestures. The table has two columns labeled 'Function' and 'Gesture 1'. Five rows show: 1) Pan - Open hand swipe gesture with illustration of flat hand moving horizontally; 2) Zoom In - Index finger and thumb moving from pinch to open position with spreading motion; 3) Zoom Out - Index finger and thumb moving from open to pinch position with closing motion; 4) Point Placement - Pinch fingers then open gesture, shown with two hand positions; 5) Switch Mode - Sweep open palm left and right, illustrated with palm moving in both directions. Each gesture is accompanied by simple black line drawings demonstrating the hand movements.

031

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 05/03/2026

New article! Anni Simola and colleagues proposed a quality assessment framework for PPGIS data and test it with three example data sets, comparing it with OpenStreetMap as reference data doi.org/10.1080/1523... #GISchat #OpenAccess

Figure 8. A map of some data from "My Green Place", a series of mapped polygons in blue overlapping each other with Green reference data behind it, covering a much larger area. Caption: Example of geometric consistency in the “My Green Place” data: mapped polygons that overlap entirely with the reference data. Background map: OpenStreetMap

011

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 04/03/2026

New article! Yue Chen and colleagues look at how we can apply continuous multi-scale transformation of building footprints that will be useful in dynamic zooming in web maps, like in the example below doi.org/10.1080/1523... See also their data at doi.org/10.6084/m9.f...

051

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 09/02/2026

New article! Zhekun Huang, Haizhong Qian and colleagues explore and evaluate new ways of matching contours from multiple sources of terrain data using unsupervised learning doi.org/10.1080/1523... Data doi.org/10.6084/m9.f... Code gitee.com/zhekunhuang/... #GISchat

Figure 12 showing visualization of contour matching results for experimental area 1 across three panels labeled b, c, and d. Each panel displays overlapping contour lines from Dataset1 (blue) and Dataset2 (orange), with cyan arrows indicating matching results identified by the proposed method. Panel b shows several circular enclosed contour features in the upper left. Panel c displays relatively parallel, wavy horizontal contours with some vertical variation. Panel d shows more complex contour patterns with enclosed shapes in the lower left and converging lines in the upper right. Each panel includes a north arrow and scale bar showing 0, 50, and 100 meters.

021

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 03/02/2026

New article! Nianhua Liu, Yu Feng and colleagues look at Trust in Climate Change Communication, including the impact of having just a map or a photo included, or just a headline #GISchat #OpenAccess doi.org/10.1080/1523... Data at figshare.com/articles/dat...

021

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 02/02/2026

New article! Manuel Riberio and colleagues use Web-Delphi and workshops to help develop map-based dashboards to support pandemic response policy-making #GISchat doi.org/10.1080/1523...

Figure 1: A horizontal diagram displaying five colorful chevron-shaped steps of a collaborative framework. Step 1 (cyan): Preparatory design of the process by decision analysis and geographic information systems experts. Step 2 (lime green): Virtual workshop with small core group to discuss Web-Delphi design. Step 3 (orange): Web-Delphi process with larger panel of experts and policymakers to ideate relevant map-based geographic information elements for policymaking in pandemic contexts. Step 4 (coral pink): Web-Delphi analyses by decision analysis and geographic information experts, including statistical analyses of results. Step 5 (purple): Virtual workshop 2 with small core group to discuss Web-Delphi results and determine which elements are relevant for policymaking in pandemic contexts.

031

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 12/01/2026

New paper! Muhammet Ali Heyik & Francisco J. Abarca-Álvarez investigate implement thick mapping in spatial design studios, providing greater understanding of complex processes and spatial understanding in complex urban environments doi.org/10.1080/1523... #GISchat

Flowchart diagram titled 'Thick mapping workflow: from collaborative site assessment to immersive installation.' The workflow progresses through four numbered stages shown in circular vignettes: (1) Introductory training and orientation phase showing people using collaborative data collection platforms like Field Maps, Ushahidi, or Emapic for faster insights; (2) Collaborative on-site assessment through geospatial data collection, depicting groups gathering information in outdoor and indoor settings; (3) Information processing through co-production by micro-groups, showing varied site observations being transformed into qualitative and quantitative metrics through asynchronous teamwork; (4) Co-creation of thick maps and installation for visual representation, illustrating how thread colors represent distinct categories (spatial, natural, or historical layers) in both a vertical hanging installation and a horizontal table-based display with suspended elements.

032

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 05/01/2026

New article! Wenbo Zhang and colleagues employ a Parameter-Efficient Fine-Tuning (PEFT) strategy to improve automated map generating using AI: MapGenerator #GISchat doi.org/10.1080/1523...

A comparison grid showing eight different map visualizations of the same geographic area featuring Silver Creek running diagonally from northwest to southeast and Interstate 65 highway running vertically to the east of the creek. The visualizations include: Label (simple map with creek and highway marked), Playground v2.5 (aerial photograph of a winding creek through green landscape), Stable diffusion 3.5 (map showing labeled creek and highway), Janus-Pro 7B (detailed street map with creek system), Map Diffusion (topographic-style map with terrain features), Flux.1-dev (minimalist map with I-65 label), GPT-4o (stylized map with creek and urban features), and MapGenerator (simplified map with creek and highway). The prompt at top describes the geographic layout with key spatial relationships highlighted in red and blue text."
The google map shows a section of a geographic area with **a creek** labeled **"Silver Creek"** running **diagonally** from the **northwest to the southeast**. To the **east** of the creek, there is a **major highway** labeled **"I-65"** running **vertically**. The background is a **light green color**, indicating land or a general area.

021

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 19/12/2025

New article! Chaun Chen, Mengyi Wei and colleagues look at GeoAI ethics and present An infographic framework of GeoAI ethics based on news data #GISchat #OpenAccess doi.org/10.1080/1523...

A horizontal stacked bar chart showing the existence percentage of different harm types in GeoAI ethics cases. Five categories are displayed with icons on the left: eco (economy), phy (physical harm), pri (violation of privacy), psy (psychological harm), and equ (violation of equal rights). Each bar is divided into two segments: magenta representing 'Existence' and tan representing 'Non-existence'. The percentages are: eco shows 41.82% existence and 58.18% non-existence; phy shows 33.22% existence and 66.78% non-existence; pri shows 30.52% existence and 69.48% non-existence; psy shows 18.90% existence and 81.10% non-existence; and equ shows the lowest at 3.60% existence and 96.40% non-existence. A legend in the upper right indicates the color coding for existence (magenta) and non-existence (tan).

031

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 18/12/2025

New article! Tomasz Opach and colleagues explore using a digital map to facilitate the exploration of place names from literary, using place names mentioned in Norwegian literature 1814–1905 doi.org/10.1080/1523... #GISchat

Three maps labeled a, b, and c showing different visualizations of the same geographic region. Map a) georeferenced map symbols, which are aggregated into proportional rectangular
map symbols, with frequencies indicated inside these shapes. Map b) displays the same region as a dot density map with hundreds of colored dots (appearing in shades of red, orange, yellow, and other colors) distributed across the area, with the highest concentration in the center. Map c) presents a heat map or kernel density visualization with colors ranging from blue (low density) through green and yellow to red (high density), showing smooth gradients of concentration with the most intense areas appearing as red hotspots in the center and upper portions of the map. All three maps include a legend, scale bar, and attribution to OpenStreetMap contributors.

041

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 15/12/2025

The final paper in our upcoming special issue on #Replicability and #Reproducibility, Joseph Holler and colleagues use open science practices to develop a GIScience study on access t oCOVID-19 healthcare in Illinois, US doi.org/10.1080/1523... #OpenAccess #GISchat

A network graph showing the evolution of seven related studies from 2020 to 2024. The timeline runs horizontally along the x-axis, with seven study names listed vertically on the left: Kang et al., CT Replication, Illinois Reproduction, Chicago Reproduction, Class Projects, VT Pharmacy Extension, and Esri Extension. Each study has a small grid showing reproducibility criteria (indicated by black and white boxes). Circles connected by lines represent different versions of each study, with solid lines indicating direct forks or pulls and dashed lines showing references. Colored symbols within or near circles indicate the type of work: green squares for reproduction, blue circles for reanalysis, red triangles for replication, and purple inverted triangles for extension. The graph shows how these studies branched, referenced, and built upon each other over time, with a note indicating that the HEGSRR template was adopted in 2022. Updates in response to reproduction efforts are noted at the top. The network demonstrates how each study reproduced, reanalyzed, replicated, or extended the original Kang et al. (2020) study (indicated by colored polygons) in sequence, improving upon different aspects of the original work (shown by black boxes in the grids).

031

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 11/12/2025

New article! How do we go about using AI to automate point symbol generation? Shuaiqing Wang, Li Shen and colleagues have a investigate, with the process and an example shown below doi.org/10.1080/1523... Data and code at doi.org/10.6084/m9.f... #GISchat

A flowchart diagram showing a two-step 'template-render' framework for map symbol generation. The process begins with a user input icon on the left. Step 1 (Template Generation) is shown in a light green box: user input and a knowledge-guided prompt flow into an LLM (represented by a robot icon), which produces a symbol description or 'template' (shown as a document icon). Step 2 (Visual Rendering) is shown in a light blue box: the symbol description flows into a T2I model (represented by a 3D cube icon), which synthesizes a visual symbol from the description and outputs a point symbol or 'render' (shown as a red map pin icon on a folded map). Arrows connect each component from left to right, illustrating the sequential workflow from user input to final rendered map symbol

A diagram illustrating point symbol concept generation using 'Zoo' as an example. On the left, a pink box shows user input: 'Zoo' or 'Design a map symbol for Zoo'. An arrow labeled 'LLM' points to a green box containing the symbol description: 'Symbol of the Zoo, abstract 2D style, white background, stylized image with animal silhouettes and tree elements, rounded rectangle shape with curved top, simple lines for texture.' An arrow labeled 'T2I model' points to the final point symbol on the right: a rounded square icon with dark green background showing silhouettes of two animals (appearing to be a lion and elephant) facing each other under a tree canopy, with a white border and rounded corners.

021

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 10/12/2025

Deepfake geography: A new paper from Valentin Meo looking at how we can detect manipulated satellite images doi.org/10.1080/1523... Also check out our interview on the paper with Valentin and @nickbearman.bsky.social: youtu.be/DklnevrHR1c #GISchat #FreeAccess

doi.org

021

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 08/12/2025

New article! Using storytelling and guided interactions to help users understand large spatial and temporal data sets. Oana Candit et al. provide an example representing active fires of 2023 Paper: doi.org/10.1080/1523... App: www.animation.oanacanditmaps.ro/app/ #GISchat

A screenshot from the app, showing a series of yellow and orange cones (each one representing fire intensity (colour) and fire frequency (width and height) along text saying "Fires are being fueled by dry conditions and strong winds".

032

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 03/12/2025

New paper! Thomas Depian et al. investigate transitions in dynamic point labelling which are an issue in dynamic mapping systems. Weather data in Vienna - the labels change and it is not obvious which have changed doi.org/10.1080/1523... #GISchat #OpenAccess doi.org/10.17605/OSF...

Figure 1. Two labelings of the weather situation in Vienna at different timestamps: 23 labels appear, 26 disappear, and 11 change their position. The labelings are computed by our prototype that we describe in Section 4. Label-Icons: © GeoSphere Austria. Note that a full visual scan of the individual labels is necessary to identify all changes (Rensink et al., 1997).

040

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 02/12/2025

New article! Timofey Samsonov explores animated transitions proportional symbol and choropleth representations on thematic maps #GISchat #ICC2023 doi.org/10.1080/1523... Check out the supplementary videos of each transition at zenodo.org/records/1717... and see it in practice at observablehq.com

Figure 2: A series of 6 maps, showing transition from proportional symbols (circles) for each area within the map to a choropleth map where the color shows the value. Labelled t=0 to t=1.

082

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 01/12/2025

We all know color in maps is important. Shangjing Jiang and colleagues used crowdsourced photographs to create place-aware colored maps which help create a sense of place doi.org/10.1080/1523... #GISchat Figure 1 below shows their comparisons: Google Maps, Snazzy Maps "Hopper" and place-aware maps

Three of the maps used in the evaluation, on the left a Google Maps style map (relatively plain colouring, with green for parks, blue for water and shades of grey and light brown for everywhere else), in the middle a Snazzy Maps - Hopper style map, with darker colors across the board, dark green for parks, dark blue for water, light green for gardens/ greenspace and grey for buildings, and on the right, the Place-aware Map, with much more vibrant colours.

031

The Cartography and Geographic Information Society @carto-giscience.bsky.social · 27/11/2025

At a festival with events happening over a week and over a whole city? Dilara Bozkurt explores temporal navigation for festival maps on mobile devices, working out how to incorporate space and time on a small screen like a mobile phone doi.org/10.1080/1523... #GISchat #OpenAccess Check out the GIF:

052