A daily scan of urban data science: what appeared, what it found, and why it matters.
7 arXiv categories, 96 journals — the 13–24 papers a day worth your time. When the signal is weak, we say so. Free, no ads, no engagement metrics.
Latest issue — 2026-08-19
Read →An anchor-sparse lane detection model for resource-constrained hardware and extreme weather
Three of today's papers carry the tag "Land Use and Ecosystem Services," focusing on stakeholder perceptions of tourism and urban growth in the Yucatan Peninsula, green space extraction and thermal comfort simulation using high-resolution remote sensing, and the 40-year relationship between urban heat islands and surface water dynamics in Gauteng Province. Outside of this group, one paper introduces SDLane, a lane detection model that uses sparse anchors and dynamic fusion to match the actual number of lane instances. Another study analyzes popular entrepreneurship as a cultural grammar based on four years of ethnographic fieldwork in the urban peripheries of São Paulo. Finally, a third paper examines how street painters in Bandung, Indonesia, perceive generative artificial intelligence and evaluates the corresponding local government policy responses.
- items
- 10
- candidates scanned
- 2405
- with code / data
- 1 / 1
- no open abstract
- 121
- tag shift
- none
Archive — last 14 days
all issues2026-08-03 quiet days in grey — each still reviews what appeared 2026-08-20
- 2026-08-20 no issue collect.openalex finished FAILED
- 2026-08-19 10 items SDLane: Efficient Lane Detection via Sparse Anchors and Dynamic Fusion 1 code, 1 data
- 2026-08-18 15 items Reintroducing a QGIS technical supplement for Mapping for a Sustainable World 1 code, 1 data
- 2026-08-14 15 items filled in later Smart routes: a system for development and comparison of algorithms for solving vehicle routing problems with realistic constraints 1 code Assembled later: nobody was watching on 2026-08-14, so this issue was built from the archived candidates afterwards.
- 2026-08-13 a quiet day filled in later Assembled later: nobody was watching on 2026-08-13, so this issue was built from the archived candidates afterwards.
- 2026-08-11 24 items RTSKG: Building a Rail Transit Station Knowledge Graph Dataset 1 data
- 2026-08-10 24 items From announcement to operation: The changing impact of light rail transit on housing prices across implementation phases 1 data
- 2026-08-09 a quiet day filled in later Assembled later: nobody was watching on 2026-08-09, so this issue was built from the archived candidates afterwards.
- 2026-08-08 16 items filled in later 3D-Geo-Vis: A Web-Based Environment for Interactive 3D Thematic Geovisualisation and Its Usability Evaluation 1 data Assembled later: nobody was watching on 2026-08-08, so this issue was built from the archived candidates afterwards.
- 2026-08-07 24 items Predicting barge tow size on inland waterways using vessel trajectory-derived features: proof of concept 2 data
- 2026-08-06 24 items Agent-based modelling of urban non-exhaust emissions reveals the limits of fleet electrification 2 data
- 2026-08-05 24 items Analyzing the daily flows: Exploring shared micro-mobility factors in Venice
- 2026-08-04 16 items filled in later Semantic enrichment of 3D city models via roof material classification for urban greening and heat island mitigation 1 code, 1 data Assembled later: nobody was watching on 2026-08-04, so this issue was built from the archived candidates afterwards.
- 2026-08-03 17 items filled in later Network-wide fine-grained passenger loading prediction via dynamic relational spatiotemporal deep graph neural networks 1 code, 1 data Assembled later: nobody was watching on 2026-08-03, so this issue was built from the archived candidates afterwards.
Method
Two entry paths. A journal article enters because a journal we track published it; an arXiv preprint enters on a classifier's probability and only above a floor set from labelled examples. Neither path is asked to fill a quota.
Honesty
Bibliography, authors and links come from the sources, never from a model. Where a measurement was not possible we leave the line out rather than print a zero, and where it was possible and found nothing we say so.
Tags
Every tag carries a canonical identifier — OpenAlex, ORCID, ROR — so a method named twice is one thing, not two. Free strings are a test failure rather than a warning.