James Williams

Web Application Development

Wed, 05 Aug 2026 10:39:24 GMT

Placeholder summary for CMM721. Replace with a real description of the module.

Information Security

Wed, 05 Aug 2026 10:39:24 GMT

Placeholder summary for CMM7XX. Replace with a real description of the module.

Introduction to Programming

Wed, 05 Aug 2026 10:39:24 GMT

Placeholder summary for CMU410. Replace with a real description of the module.

Web Applications

Wed, 05 Aug 2026 10:39:24 GMT

Placeholder summary for CMU421. Replace with a real description of the module.

Fundamentals of Web Design

Wed, 05 Aug 2026 10:39:24 GMT

Placeholder summary for CMU422. Replace with a real description of the module.

Fundamentals of Systems and Networking

Wed, 05 Aug 2026 10:39:24 GMT

Placeholder summary for CMU4XX. Replace with a real description of the module.

Advanced Web Development

Wed, 05 Aug 2026 10:39:24 GMT

Placeholder summary for CMU529. Replace with a real description of the module.

Cyber Security

Wed, 05 Aug 2026 10:39:24 GMT

Placeholder summary for CMU540. Replace with a real description of the module.

Industry Project

Wed, 05 Aug 2026 10:39:24 GMT

Placeholder summary for CMU597. Replace with a real description of the module.

Software Security

Wed, 05 Aug 2026 10:39:24 GMT

Placeholder summary for CMU5XX. Replace with a real description of the module.

Information Security for Industry

Wed, 05 Aug 2026 10:39:24 GMT

Placeholder summary for CMU5XX-INFOSEC. Replace with a real description of the module.

Dissertation

Wed, 05 Aug 2026 10:39:24 GMT

Placeholder summary for CMU601. Replace with a real description of the module.

Ethical Hacking

Wed, 05 Aug 2026 10:39:24 GMT

Placeholder summary for CMU6XX. Replace with a real description of the module.

Cyber Forensics

Wed, 05 Aug 2026 10:39:24 GMT

Placeholder summary for CMU6XX-CF. Replace with a real description of the module.

Applied Electrical and Electronic Engineering: Construction Project

Wed, 05 Aug 2026 10:39:24 GMT

Placeholder summary for EEEE1002. Replace with a real description of the module.

Introduction to Software Engineering and Programming

Wed, 05 Aug 2026 10:39:24 GMT

Placeholder summary for EEEE1040. Replace with a real description of the module.

HDL for Programmable Logic

Wed, 05 Aug 2026 10:39:24 GMT

Placeholder summary for EEEE4123. Replace with a real description of the module.

General capability barely predicts spatial reasoning in LLMs

Wed, 05 Aug 2026 00:00:00 GMT

I took the eleven top models on the Vellum leaderboard and asked them 46 spatial questions with machine-computed answers: GeoJSON axis order, great circle distance and bearing, UTM zones including the Norway and Svalbard exceptions, RFC 7946 winding order, and point-in-polygon on a concave ring. One run, temperature 0, through OpenRouter. The correlation between general capability and spatial score is **r = 0.19**. | Model | Score | Answered | Of answered | Mean tokens | Vellum | |---|---:|---:|---:|---:|---:| | GPT-5.6 Sol | 97.6 | 97.6 | 100.0 | 711 | 47.2 | | Claude Opus 5 | 95.7 | 100.0 | 95.7 | 532 | 64.7 | | Claude Fable 5 | 91.3 | 100.0 | 91.3 | 415 | — | | Claude Sonnet 5 | 91.3 | 91.3 | 100.0 | 844 | 57.4 | | Claude Opus 4.8 | 69.6 | 100.0 | 69.6 | 37 | 57.9 | | DeepSeek V4 Flash | 69.6 | 100.0 | 69.6 | 166 | 51.6 | | GLM 5.2 | 63.0 | 69.6 | 90.6 | 1530 | 54.7 | | DeepSeek V4 Pro | 56.5 | 60.9 | 92.9 | 2104 | 48.2 | | Kimi K3 | 45.7 | 67.4 | 67.7 | 1715 | 56.0 | | Kimi K2.6 | 28.3 | 30.4 | 92.9 | 3128 | 54.0 | | Gemini 3.1 Pro | 26.1 | 100.0 | 26.1 | 195 | — | The lowest-ranked model on general capability came first. The highest-ranked came second. Below that the ordering scrambles entirely. Three columns matter more than the score. **Answered** is the share that produced anything within a 4,000-token budget. **Of answered** is accuracy on those. **Mean tokens** shows why they differ: Kimi K2.6 spends 3,128 tokens per question and answers 30% of them, but is right 93% of the time when it finishes. It is not bad at geodesy, it runs out of room. Reporting those as one number is a trap I fell into first time round. Capped at 200 tokens, ten of eleven models scored zero on distance. They had not failed, they had been truncated, and an empty response was being scored as a wrong answer. The resulting leaderboard ranked models by terseness and looked entirely plausible. Axis order is near-solved: nine of eleven score 100. Bearing is not: four score zero. Nobody should be computing geodesy in a language model anyway, but plenty of [agent frameworks now do](/work/geon). It is also [why I assess reading code](/notes/teaching-web-development-when-code-is-free). *46 tasks, one run, $5.20. Ground truth computed, not written. Single run, so no variance estimate.*

AnythingPOI: a fused, confidence-scored points of interest dataset

Tue, 04 Aug 2026 00:00:00 GMT

OpenStreetMap and Overture Maps conflated into one deduplicated, classified, openly licensed dataset. 22.7M points across six countries, every record carrying a confidence score and the evidence behind it.

CDISAW: a queryable data infrastructure for slavery and armed conflict

Tue, 04 Aug 2026 00:00:00 GMT

A research platform that makes 53 incompatible datasets on slavery in war answerable as one corpus. 23.5M linked event records across five ontological layers, queried through a typology-first interface, with every query citable.

GEON: an LLM-native format for representing places as meaning, not coordinates

Tue, 04 Aug 2026 00:00:00 GMT

An open text format that describes what a place is, how it works and how it feels, rather than where its vertices are. Three reference implementations, bidirectional GeoJSON conversion, and measurably fewer tokens than the format it complements.

Leisure Walking Systems Working Group: making place research buildable

Tue, 04 Aug 2026 00:00:00 GMT

An EPSRC-funded academic and industry initiative that produced open technical documentation, design frameworks and evaluation tools for place-based walking systems. 25+ documents across 8 research areas, published for practitioners rather than archived for academics.

PARM-X: projecting live geospatial data onto physical city models

Tue, 04 Aug 2026 00:00:00 GMT

The interactive software layer for Nottingham's projection augmented relief model. Live council data, crowd simulations and public responses draped onto physical terrain across four synchronised screens, cutting a four-hour preparation process to minutes for executive planning sessions.

Topodex: contextual geocoding for conflict and human rights corpora

Tue, 04 Aug 2026 00:00:00 GMT

A Python library that makes geocoders accountable to the document they are reading. 894M+ places resolved across nine open sources, reranked against document context, with a six-category failure taxonomy and a coherence check over everywhere a document names.

WalkGIS: linking video narratives to maps to capture sense of place

Tue, 04 Aug 2026 00:00:00 GMT

A contextual GIS that fuses first-person walk video, think-aloud narration and GPS into one synchronised environment, so what a walker said can be queried spatially. Analysis per walk fell from two or three weeks to one.

WalkGrid: personalising urban walking through environmental similarity

Tue, 04 Aug 2026 00:00:00 GMT

A routing platform that curates walks by what makes them worth taking, from greenspace and heritage to air quality and safety, rather than by distance. 51 environmental features scored across a hexagonal grid, with route selection driven by natural language.

The curse of the .science domain: why this site moved to jameswil.com

Sat, 01 Aug 2026 00:00:00 GMT

This site used to live at jwilliams.science. It now lives at jameswil.com, because a domain that some readers cannot open is not a domain. The reason is the TLD, not the site. `.science` was cheap at launch and got used accordingly, so it carries a poor reputation with the blocklists that corporate proxies, university filters and mail gateways buy in wholesale. Those lists work at the TLD level. Nothing on my site was ever assessed. Readers on several Canadian networks simply got a block page, and a CV link that returns a block page is worse than no link. Mail is the sharper edge. A `.science` address in a From header is enough to score a message into a junk folder before anything else about it is read, which matters when the message is an application. The lesson is dull and worth stating: a novelty TLD is not a saving, it is a deliverability decision made once and paid for repeatedly. Old links redirect, and the five references to the old domain still sitting in [project](/work/anythingpoi) [pages](/work/lwswg) are next.

Teaching web development when writing the code is the cheap part

Tue, 14 Jul 2026 00:00:00 GMT

A first-year student can now produce a working CRUD app in an afternoon. Assessing the artefact stopped being informative the moment that became true, because the artefact no longer evidences anything about the student. What has not become cheap is diagnosis. Generated code fails in ordinary ways: a fetch that never handles the rejected promise, an auth check enforced in the client and nowhere else, an N+1 query that is invisible at 20 rows and fatal at 20,000. A student who cannot read the code cannot see any of it, and the model will not volunteer it. It is the same failure I measure elsewhere: [confident output, no signal about whether it is right](/notes/general-capability-does-not-predict-spatial). So the marks moved. On [CMU422](/teaching/cmu422) and [CMU529](/teaching/cmu529) the submission is still an application, but it arrives as a repository: commit history, a README that justifies the dependencies and the architecture, and a short written reflection on what broke, what the student changed their mind about, and what they still do not understand. A repo you cannot explain is a repo you did not build. This is not a restriction on tool use, and policing that would fail anyway. It is a change in what counts as the work. Reading code critically was always the harder skill and we could previously get away with assuming it followed from writing enough of it. It does not follow any more, so it has to be taught directly.

POI conflation thresholds: why a 2% cross-source match rate is the honest answer

Tue, 23 Jun 2026 00:00:00 GMT

Two POIs sit 30 m apart with similar names. Same shop, or two shops? Every conflation pipeline answers that question and almost none of them publish the answer. [AnythingPOI](/work/anythingpoi), which fuses OpenStreetMap and Overture Maps into 22.7M points across six countries, uses three gates, all of which must hold: 50 m Haversine, agreement on Tier-1 category, and Jaro-Winkler name similarity of 0.85 on normalised names. Candidates are blocked first by [H3 at resolution 11](/notes/h3-kring-is-not-the-radius-you-asked-for), which is the same parameter choice in a different costume. Across six countries the gates match 1.4% to 3.2% of records. The number looks like failure. Loosening it is trivial: take the gate to 150 m, drop the category check, and the merged count climbs to something that reads well in a table. What you have actually done is fuse the barber and the newsagent that share a parade and a family name, and you cannot tell which of your merges are real. A cross-source match is worth something in the confidence score precisely because it is rare. It means two independently built datasets agree. Inflate the rate and you destroy the signal that made the field worth adding. Recall is what this costs. Transliterated names go unmatched, and so does a shopping centre whose centroid sits 200 m from its entrance. That is a deliberate trade, and it belongs in the documentation rather than in a footnote.

Micro visualisations for spatial data: draw the geometry, not the number

Tue, 21 Apr 2026 00:00:00 GMT

A table of metres cannot say what a neighbourhood is shaped like. The [k-ring that overshoots 400 m](/notes/h3-kring-is-not-the-radius-you-asked-for) reaches furthest along six axes, so a shop 390 m away is inside or outside depending on its bearing. Inline SVG draws that at the size of the text, inside the sentence making the claim. Resolution 9 gives you for a 400 m walk, and resolution 11 gives you . Same dashed 400 m target, same 15 px, and only the second is the neighbourhood you asked for. Overshoot falls from 514% to nothing. At 64 px there is room for the boundary. Real H3 disks at 51.48 degrees north, each scaled to its own footprint: | Res | Footprint | Reach | Overshoot | Cells | |---:|---|---:|---:|---:| | 7 | | 2,457 m | +514% | 7 | | 8 | | 922 m | +130% | 7 | | 9 | | 702 m | +75% | 19 | | 10 | | 527 m | +32% | 61 | | 11 | | 401 m | +0% | 217 | | 12 | | 414 m | +3% | 1,519 | Read the dashed circle, not the outline. At resolution 7 the target is a speck inside the one cell you were given. By 11 it nearly fills a footprint that has stopped being a hexagon, which is what 217 cells bought. Each is under a kilobyte, needs no charting library, and has fixed `width` and `height`, so it cannot shift the layout. `currentColor` serves both themes from one file. Best of all, the script that computes the table emits the glyphs, so the picture cannot drift from the numbers beside it. A figure does not have to be a page wide to be a figure.

Choosing an H3 resolution for neighbourhood analysis: your k-ring is not the radius you asked for

Tue, 17 Feb 2026 00:00:00 GMT

Ask for a [400 m walkable neighbourhood](/work/walkgrid), the standard five-minute walk, and H3 hands you something else. At resolution 9 you get 702 m. That is not a rounding error, it is a neighbourhood 76% too big. The cause is that `k` is an integer and the grid step is fixed. A disk grows in whole rings, so its radius is wherever the grid happens to land, never the number you asked for. Coarse resolutions land badly. Grown rather than estimated, at 51.48°N: ```python k = 1 while disk_reach(cell, k) < 400: k += 1 ``` Resolution 9 reaches 702 m for 19 cells, resolution 10 reaches 527 m for 61, and only [resolution 11](/notes/poi-conflation-match-rate), the resolution [AnythingPOI](/work/anythingpoi) blocks on, lands at 401 m, for 217. Measuring the disk a second way, as the radius of the circle of equal area, agrees: 769 m, 521 m, 371 m. So each step costs roughly 3.5× the cells, and what it buys is boundary accuracy. It is not buying finer cells, which mostly encode precision your input data does not have. Pick the resolution from the error you will accept on the boundary, then check the cell count you can afford. Choosing on cell size alone is how a 400 m study quietly becomes a 700 m one. Cells are not equal-area, so these are numbers for 51°N. Rerun them at yours.

Multi-Resolution H3 Walkability Surfaces: A Scalable Framework for Cross-City Pedestrian Environment Assessment

Thu, 01 Jan 2026 00:00:00 GMT

Introduction: Quantifying urban walkability at fine spatial resolution is essential for evidence-based active travel planning. Yet, existing indices typically rely on coarse administrative units or proprietary data, which impedes cross-city comparisons. This study presents the Hexagonal Walkability Surface (HWS) framework: a configuration-driven pipeline for computing multi-resolution walkability indices from geospatial data. Methods: HWS employs the H3 discrete global grid to tessellate city extents at four hierarchical resolutions (levels 8-11). Five pedestrian-environment features are extracted per hexagon: (1) footway and path density (pedestrian edge length per km²); (2) point-of-interest accessibility, combining in-cell counts with an inverse-distance-weighted proximity score; (3) land-use diversity as Shannon entropy of intersecting land-use classes; (4) greenspace proximity, integrating coverage proportion with distance decay; and (5) road-safety ratio of pedestrian-friendly to total edge length. Street network and land-use data are sourced from OpenStreetMap; place data is augmented with Overture Maps (2026-02-18) via DuckDB S3 queries. Features are min-max normalised and combined as a weighted composite score (0-100). Spatial autocorrelation (Getis-Ord Gi*) is used to identify clusters of walkability. Results: In Nottingham and Bristol, United Kingdom, the framework produces 1,178-57,700 and 2,581-126,518 hexagons per city, respectively, across resolutions 9-11. Across the resolutions, agreement is high (Spearman ρ = 0.963-0.971 between adjacent levels, all p < 0.001), confirming multi-scale consistency. At resolution 9 (roughly neighbourhood scale), intra-urban variation is observed in both cities (σ = 11.4 and 10.1 score points respectively), with city-centre hexagons clustering as hotspots and peripheral areas as coldspots. Gini coefficients indicate greater inequality in walkability in Nottingham (0.36) than in Bristol (0.26). Conclusions: HWS provides a reproducible, data-agnostic methodology extensible to any city with OSM coverage. The multi-resolution design could support both strategic (neighbourhood) and street-block planning decisions. By making fine-grained walkability assessment accessible to researchers and planners without proprietary data or bespoke infrastructure, HWS lowers the barrier to evidence-based active travel policy at the urban scale.

Macro-Regional Spatial Patterns of Ambient Air Pollution and Avoidable Hospitalizations for Community-Acquired Pneumonia in Mexico (2013–2020)

Thu, 01 Jan 2026 00:00:00 GMT

Ambient air pollution significantly contributes to respiratory illnesses, yet little is known about how industrial emissions are linked to preventable hospitalizations across atmospheric basins in middle-income countries. This study develops a basin-based geo-matics framework to examine the spatial and temporal relationship between industrial pollutants and age- and sex-adjusted avoidable hospitalizations for community-acquired pneumonia (PQI 11) in Mexico from 2013 to 2020. Using state-level data grouped into eight macro-regions, we combine bivariate choropleth maps, Pearson correlations, linear regression, and longitudinal time-series analysis to identify spatial clusters of high risk and to estimate regional sensitivities to changes in PM2.5, SO2, NOx, and volatile organic compound emissions. The findings reveal notable regional differences: northern border states and the Mexico City metropolitan basin form persistent high–high clusters where elevated emissions coincide with high PQI 11 rates, while coastal and peninsular regions show lower hospitalization burdens despite medium emission levels. Although national industrial PM2.5 emissions decreased over the study period, several macro-regions—particularly CDMX_Edomex, Centro, and Centro Norte—experienced significant increases in avoidable hospitalizations and decoupled emission–health patterns. Correlation matrices and regression slopes suggest that the strength and even direction of links between pollutants and PQI 11 vary across macro-regions, with emission-responsive patterns in Centro Norte and weak or inverse relationships in Peninsula and Pacifico Sur. These findings demonstrate that national averages obscure critical spatial disparities and highlight the value of basin-based geomatics approaches for regional air-quality governance, spatial decision support, and primary-care planning aimed at reducing preventable respiratory hospitalizations.

Geospatial Experience-Oriented Notation (GEON): A Semantic Format for LLM-Native Spatial Intelligence

Thu, 01 Jan 2026 00:00:00 GMT

Existing geospatial data formats such as GeoJSON, Well-Known Text (WKT), and CityGML, are optimised for geometric computation and rendering. While effective for Geographic Information Systems (GIS), these approaches present limitations when used with Large Language Models (LLMs). For example, coordinate arrays carry no inherent semantic meaning, spatial relationships require computational geometry to extract, and the human experience of place is usually absent. This manuscript introduces foundational work on Geospatial Experience-Oriented Notation (GEON), a text-based format that bridges machine-optimised geospatial data and human-readable spatial descriptors. GEON encodes identity, geometry, purpose, experiential qualities, spatial relationships, temporal patterns, and data provenance in a readable and structured syntax designed for human comprehension and LLM reasoning. This manuscript presents the initial specification, reference implementations in Python, Rust, and JavaScript, and an empirical evaluation demonstrating how GEON achieves 20% fewer tokens than equivalent GeoJSON files, while encoding 31% more semantic facts per token. This manuscript explores the implementation and how LLMs reason about place-making, urban design interventions, and spatial intelligence tasks that existing formats struggle to support.

AnythingPOI - Australia POI Dataset v0.1

Thu, 01 Jan 2026 00:00:00 GMT

A unified, open point-of-interest (POI) dataset for Australia containing 1,735,980 POIs produced by the AnythingPOI pipeline, which fuses OpenStreetMap and Overture Maps Foundation data using H3-indexed spatial conflation, Jaro-Winkler name matching, and multi-signal confidence scoring. Source breakdown: OSM-only: 320,149 (18.4%) — from OpenStreetMap contributors (ODbL 1.0) Overture-only: 1,364,195 (78.6%) — from Overture Maps Foundation (CDLA-Permissive-2.0) Conflated (both sources matched): 51,636 (3.0%) Top categories: Professional & Business, Retail, Food & Beverage, Transportation, Healthcare. Full taxonomy: 18 Tier-1 categories, 196 Tier-2 subcategories. Contents: GeoParquet files (one per Tier-1 category), PMTiles v3 for interactive map visualisation, and coverage statistics CSVs. Each POI carries a confidence_score (0.01–0.99) reflecting the strength of the conflation evidence across spatial, name, website, phone, postcode, and Wikidata signals. Attribution: This dataset contains information from OpenStreetMap (© OpenStreetMap contributors, ODbL 1.0 — openstreetmap.org/copyright) and Overture Maps Foundation (CDLA-Permissive-2.0 — overturemaps.org). License: Open Database License (ODbL) 1.0. Any public use of this database or works produced from it must include the above attribution. Derivative databases must also be released under ODbL.

AnythingPOI - Canada POI Dataset v0.1

Thu, 01 Jan 2026 00:00:00 GMT

A unified, open point-of-interest (POI) dataset for Canada containing 5,565,256 POIs produced by the AnythingPOI pipeline, which fuses OpenStreetMap and Overture Maps Foundation data using H3-indexed spatial conflation, Jaro-Winkler name matching, and multi-signal confidence scoring. Source breakdown: OSM-only: 451,872 (8.1%) — from OpenStreetMap contributors (ODbL 1.0) Overture-only: 5,037,979 (90.5%) — from Overture Maps Foundation (CDLA-Permissive-2.0) Conflated (both sources matched): 75,405 (1.4%) Top categories: Professional & Business, Retail, Food & Beverage, Healthcare, Services. Full taxonomy: 18 Tier-1 categories, 196 Tier-2 subcategories. Contents: GeoParquet files (one per Tier-1 category), PMTiles v3 for interactive map visualisation, and coverage statistics CSVs. Each POI carries a confidence_score (0.01–0.99) reflecting the strength of the conflation evidence across spatial, name, website, phone, postcode, and Wikidata signals. Attribution: This dataset contains information from OpenStreetMap (© OpenStreetMap contributors, ODbL 1.0 — openstreetmap.org/copyright) and Overture Maps Foundation (CDLA-Permissive-2.0 — overturemaps.org). License: Open Database License (ODbL) 1.0. Any public use of this database or works produced from it must include the above attribution. Derivative databases must also be released under ODbL.

AnythingPOI - Germany POI Dataset v0.1

Thu, 01 Jan 2026 00:00:00 GMT

A unified, open point-of-interest (POI) dataset for Germany containing 6,763,796 POIs produced by the AnythingPOI pipeline, which fuses OpenStreetMap and Overture Maps Foundation data using H3-indexed spatial conflation, Jaro-Winkler name matching, and multi-signal confidence scoring. Source breakdown: OSM-only: 2,420,344 (35.8%) — from OpenStreetMap contributors (ODbL 1.0) Overture-only: 4,129,920 (61.1%) — from Overture Maps Foundation (CDLA-Permissive-2.0) Conflated (both sources matched): 213,532 (3.2%) Top categories: Professional & Business, Transportation, Retail, Food & Beverage, Other / Uncategorized. Full taxonomy: 18 Tier-1 categories, 196 Tier-2 subcategories. Contents: GeoParquet files (one per Tier-1 category), PMTiles v3 for interactive map visualisation, and coverage statistics CSVs. Each POI carries a confidence_score (0.01–0.99) reflecting the strength of the conflation evidence across spatial, name, website, phone, postcode, and Wikidata signals. Attribution: This dataset contains information from OpenStreetMap (© OpenStreetMap contributors, ODbL 1.0 — openstreetmap.org/copyright) and Overture Maps Foundation (CDLA-Permissive-2.0 — overturemaps.org). License: Open Database License (ODbL) 1.0. Any public use of this database or works produced from it must include the above attribution. Derivative databases must also be released under ODbL.

AnythingPOI - Netherlands POI Dataset v0.1

Thu, 01 Jan 2026 00:00:00 GMT

A unified, open point-of-interest (POI) dataset for Netherlands containing 1,782,538 POIs produced by the AnythingPOI pipeline, which fuses OpenStreetMap and Overture Maps Foundation data using H3-indexed spatial conflation, Jaro-Winkler name matching, and multi-signal confidence scoring. Source breakdown: OSM-only: 280,707 (15.7%) — from OpenStreetMap contributors (ODbL 1.0) Overture-only: 1,451,063 (81.4%) — from Overture Maps Foundation (CDLA-Permissive-2.0) Conflated (both sources matched): 50,768 (2.8%) Top categories: Professional & Business, Retail, Food & Beverage, Services, Healthcare. Full taxonomy: 18 Tier-1 categories, 196 Tier-2 subcategories. Contents: GeoParquet files (one per Tier-1 category), PMTiles v3 for interactive map visualisation, and coverage statistics CSVs. Each POI carries a confidence_score (0.01–0.99) reflecting the strength of the conflation evidence across spatial, name, website, phone, postcode, and Wikidata signals. Attribution: This dataset contains information from OpenStreetMap (© OpenStreetMap contributors, ODbL 1.0 — openstreetmap.org/copyright) and Overture Maps Foundation (CDLA-Permissive-2.0 — overturemaps.org). License: Open Database License (ODbL) 1.0. Any public use of this database or works produced from it must include the above attribution. Derivative databases must also be released under ODbL.

AnythingPOI - Türkiye POI Dataset v0.1

Thu, 01 Jan 2026 00:00:00 GMT

A unified, open point-of-interest (POI) dataset for Türkiye containing 2,201,304 POIs produced by the AnythingPOI pipeline, which fuses OpenStreetMap and Overture Maps Foundation data using H3-indexed spatial conflation, Jaro-Winkler name matching, and multi-signal confidence scoring. Source breakdown: OSM-only: 306,263 (13.9%) — from OpenStreetMap contributors (ODbL 1.0) Overture-only: 1,864,146 (84.7%) — from Overture Maps Foundation (CDLA-Permissive-2.0) Conflated (both sources matched): 30,895 (1.4%) Top categories: Retail, Professional & Business, Food & Beverage, Services, Community. Full taxonomy: 18 Tier-1 categories, 196 Tier-2 subcategories. Contents: GeoParquet files (one per Tier-1 category), PMTiles v3 for interactive map visualisation, and coverage statistics CSVs. Each POI carries a confidence_score (0.01–0.99) reflecting the strength of the conflation evidence across spatial, name, website, phone, postcode, and Wikidata signals. Attribution: This dataset contains information from OpenStreetMap (© OpenStreetMap contributors, ODbL 1.0 — openstreetmap.org/copyright) and Overture Maps Foundation (CDLA-Permissive-2.0 — overturemaps.org). License: Open Database License (ODbL) 1.0. Any public use of this database or works produced from it must include the above attribution. Derivative databases must also be released under ODbL.

AnythingPOI - United Kingdom POI Dataset v0.1

Thu, 01 Jan 2026 00:00:00 GMT

A unified, open point-of-interest (POI) dataset for United Kingdom containing 4,622,174 POIs produced by the AnythingPOI pipeline, which fuses OpenStreetMap and Overture Maps Foundation data using H3-indexed spatial conflation, Jaro-Winkler name matching, and multi-signal confidence scoring. Source breakdown: OSM-only: 947,908 (20.5%) — from OpenStreetMap contributors (ODbL 1.0) Overture-only: 3,551,749 (76.8%) — from Overture Maps Foundation (CDLA-Permissive-2.0) Conflated (both sources matched): 122,517 (2.7%) Top categories: Professional & Business, Retail, Food & Beverage, Other / Uncategorized, Transportation. Full taxonomy: 18 Tier-1 categories, 196 Tier-2 subcategories. Contents: GeoParquet files (one per Tier-1 category), PMTiles v3 for interactive map visualisation, and coverage statistics CSVs. Each POI carries a confidence_score (0.01–0.99) reflecting the strength of the conflation evidence across spatial, name, website, phone, postcode, and Wikidata signals. Attribution: This dataset contains information from OpenStreetMap (© OpenStreetMap contributors, ODbL 1.0 — openstreetmap.org/copyright) and Overture Maps Foundation (CDLA-Permissive-2.0 — overturemaps.org). License: Open Database License (ODbL) 1.0. Any public use of this database or works produced from it must include the above attribution. Derivative databases must also be released under ODbL.

A Framework for Curating Personalised Leisure Walking Experiences

Wed, 01 Jan 2025 00:00:00 GMT

How can a richer understanding of the leisure walking experience be used to support the curation of personalised route recommendations? Leisure walking is a personal and subjective experience that encompasses a range of multi-faceted expectations and narratives, this can include visiting points of interests, connecting with the environment, or engaging with the social fabric of places. The broad and disparate scope of these reasons and interests makes the process of recommending new and personalised leisure walking experiences difficult. Existing research exploring the recommendation of leisure walking experiences is often based on broad assumptions about walkers with little representation of subjective or contextual detail. Prior work in leisure walking fails to address the wide array of reasons for leisure walking and in turn representing these in personalised walking experiences. Based on the lack of personalisation of leisure walking experiences, this thesis investigates leisure walking from a user-centred perspective. Three grounded theory studies are conducted to understand leisure walking, capturing details on (1) leisure walking behaviours through a behaviour survey, (2) practitioner knowledge of the subject area through interviews with professionals, and (3) a rich understanding of the leisure walking experience through a think-aloud study. Grounded theory is used in this thesis to address the broad assumptions about walking, developing a theoretical understanding of leisure walking grounded in empirical studies. Using this grounded theory of leisure walking behaviours, professional perspectives, and walkers in-situ experiences, a framework is designed to support the curation of personalised leisure walking experiences. The framework represents the research related to three tasks of leisure walking: planning, doing, and reflecting. Using this understanding a demonstrator tool for curating personalised leisure walking experiences is designed based on forty-nine properties and considerations formed from the grounded theory. A think-aloud and in-depth interview study is conducted to evaluate the role of the tool in supporting the curation of personalised experiences based on the participants local knowledge of an area. The qualitative evaluation found that the system is able perform well in terms of matching local knowledge and supporting the curation of new experiences, often recommending routes that can either be explained by the participant or which match expectations. The thesis closes with a discussion on future opportunities for leisure walking technology, providing design considerations for supporting personalised leisure walking experiences.

Diabetes Disparities in Mexico: A Spatio-Temporal and Marginalization Index Analysis

Wed, 01 Jan 2025 00:00:00 GMT

Understanding the geospatial and temporal distribution of diabetes mellitus in Mexico can be an essential tool in supporting vulnerable populations and addressing health inequalities. This article presents a spatio-temporal investigation of patients aged 18 years and older with diabetes mellitus in Mexico, associated with geographical area and a temporal range from 2005 to 2022. This approach includes calculating diabetes-related hospitalizations and deaths and its association with the margination index segmented into eight geographical areas of Mexico. Furthermore, this research stratifies based upon age group and type of medical institute of the health services in Mexico. The main contribution of this research is to explore the relationship between diabetes-related hospitalizations, deaths, geographical area, age, sex, and margination index of populations to support preventive action. The results highlight that adults between the ages of 45 and 64 years old who live in areas with a high margination index have a greater likelihood of suffering complications related to diabetes. The age-adjusted rate of DRAH shows that the Peninsula has the highest values among geographical areas. Research will now continue to explore mapping interventions to specific states and external datasets, to further extrapolate the results of the analysis.

PlaceAgents: Multi-Stop Pedestrian Itineraries as Platial Flows on Urban Networks

Wed, 01 Jan 2025 00:00:00 GMT

This article presents PlaceAgents, a framework for modelling multi-stop pedestrian itineraries on city networks using open and reproducible data. The approach uses the itinerary as the base behavioural unit and links three core principles: (1) access to open data sources, (2) enabling explicit assumptions about agents, and (3) the auditability of outputs.

PlaceCrafter: Curating Urban Functional Regions through Platial Clustering of OpenStreetMap Points of Interest

Wed, 01 Jan 2025 00:00:00 GMT

The world is not just made of streets, buildings, and zones; it is shaped by how people engage and interact with places in their everyday lives. This manuscript presents a web-based geospatial tool that enables the mapping of these lived places and locales named PlaceCrafter. PlaceCrafter supports researchers in identifying platial regions: functional, human-centred areas that cross administrative and formal boundaries. The framework is built on OpenStreetMap, combining (near) real-time clustering, analysis, and statistical validation of these platial regions. PlaceCrafter is an initial demonstration for exploring the subjective experiences of place through existing datasets and city structures.

Towards a Framework for Personalising Leisure Walking Route Recommendations

Mon, 01 Jan 2024 00:00:00 GMT

This research investigates how a greater understanding of leisure walking can be used to support the personalisation and curation of new leisure walking experiences. Existing solutions are often limited in the range of routing properties a user has access to. The purpose of this research is to explore a richer understanding of how, what, why, and where leisure walkers engage with walks. Through a grounded theory approach combining a walking behaviour survey, a think-aloud study, and an expert interview study, a framework for personalising leisure walking route recommendations has been designed. The remaining work includes finalising the development of a web-based GIS demonstrator system for curating personalised routes and conducting an evaluation of this approach.

Emerging Platial Narratives and Themes from a Leisure Walking Study

Sun, 01 Jan 2023 00:00:00 GMT

This article presents the preliminary results of a think-aloud leisure walking study, identifying the key themes and platial narratives. A think-aloud study was conducted to explore what and how leisure walkers engaged with while walking. Our emerging results are presented in the context of an approach to extracting and understanding the platial experience during the study. The early findings suggest that the types of places engaged with while walking and the characteristics of these places are varied, while navigation and wayfinding have an impact on the selected route and the changes that occur during the walk. Our future work will now focus on further analysing these results and using them to improve the recommendation of leisure walking routes.

Survey of leisure walking behaviours and activity tracking use

Sun, 01 Jan 2023 00:00:00 GMT

Williams, J., Pinchin, J., Hazzard, A. & Priestnall, G. (2023). Survey of leisure walking behaviours and activity tracking use. Coordinates Magazine, 13-15. Coordinates.

WalkGIS: Exploring Platial Analysis of Leisure Walks via Linked Video Narratives

Sun, 01 Jan 2023 00:00:00 GMT

Extracting rich contextual information from study participants presents an interesting challenge when the expected results are uncertain. This article presents the design of a contextual geographic information system (GIS) to extract platial information from a multimodal data set (audio, video, and GPS) collected during a ‘think-aloud’ leisure walking study. WalkGIS enables transcriptions, labelling, and platial analysis to be performed within one system, with data being linked and coordinated to form linked video narratives.

An Emerging Conceptual Model for Curating Engaging Leisure Walking Recommendations

Sat, 01 Jan 2022 00:00:00 GMT

Providing routes to leisure walkers requires alternative recommendation scenarios to those used in tourism routing systems. In this paper, we present an emerging conceptual model of three scenarios for curating leisure walking route recommendations. Our recommendation scenarios consider the highest ranked similar walks, routes for new application users, and a progressively changing route recommendation scenario. Conceptual models for these scenarios are presented and the challenges in completing this research are considered. Feedback received on these early conceptual models will be used to further design a recommendation framework for curating engaging leisure walking experiences.

Context for Leisure Walking Routes: A Vision for a Spatial-Platial Approach

Sat, 01 Jan 2022 00:00:00 GMT

Providing recommendations for interesting and engaging leisure walking routes is a complex problem due to the subjective and personal nature of the activity. Existing work has often focused on recommending the quickest or most popular walks. However, these routes often lack detail on the contextual and experiential factors of walks and do not attempt to match the requirements with those of users. This article presents a vision of how more contextual detail can be applied to walking routes. We consider how existing analysis and spatial data mining techniques, including real-time clustering, viewshed analysis, and colocation patterns, could be used to extend a place-based understanding of leisure walking routes. By using spatial methods to extrapolate a rich platial understanding of the locations of a walk, the proposed methods in this article will support an emerging framework for curating engaging leisure walking experiences, recommending routes beyond those of the quickest or the most popular.

CrowdTracing: Overcrowding Clustering and Detection System for Social Distancing

Fri, 01 Jan 2021 00:00:00 GMT

Kanjo, E., Anderez, D. O., Anwar, A., Al Shami, A. & Williams, J. (2021). CrowdTracing: Overcrowding Clustering and Detection System for Social Distancing. 2021 IEEE International Smart Cities Conference (ISC2), 1-7. https://doi.org/10.1109/ISC253183.2021.9562914

Survey of Leisure Walking Behaviours and Activity Tracking Use: Emerging Themes and Design Considerations

Fri, 01 Jan 2021 00:00:00 GMT

In this paper we present a work in progress analysis of a leisure walking behaviour survey that focuses on walkers' habits and experiences. We are specifically interested in the use of mobile tracking applications in this context to help design and deploy future technologies that can better support engaging leisure walks through synthesising previous behaviours and experiences. This survey collected 329 responses relating to self-reported walking behaviour patterns and mobile activity tracker use. In the emerging analysis we identified design considerations for future walking-focused applications, emphasizing the subjective and personal nature of walking routes.

FinVis: Visualizing the Complex Nature of Financial Markets

Wed, 01 Jan 2020 00:00:00 GMT

Investors are constantly looking for insights through comparisons on stock market data to assist in the discovery of well-priced stocks. Modern stock market applications and research allow for a range of visualizations to be produced whilst enabling the viewing of complete datasets, however, these tools are often split, not providing queries and analysis in the same view, therefore, limiting the potential of the proposed solution. Within this thesis, a new system for performing analysis and comparison is proposed, a web-based application which is able to assist in the exploratory visual analysis and visualization generation of financial data. FinVis allows for over 500 stocks to be queried and 10,000 to be imported, before displaying interactive and customizable visualizations to assist in the discovery of bargains or good investments on the market. The application includes tools to save, store and load queries alongside a visual analytic display to enable customization and more complex individual views to be generated. A complete set of interactions are also provided to the user, whilst enabling all of this exploration to take place with real-time and historical data being provided through a financial API to the user’s web browser. The thesis enables future work within the area of exploratory financial visualization to take place, providing core concepts in a comparative nature to potential researchers and investors alike.