Constarium
← Search

Table · dataset · 2026

ShadeWalk data: shadow-model rasters, covered-linkway training tiles, transit and building layers for Singapore

Listed in figshare

Description

<p dir="ltr">Input layers of the ShadeWalk pipeline (shade-facility extraction, facility-aware SOLWEIG-GPU shadow modelling, shade-aware pedestrian-network reconstruction and coolest-route pedestrian-flow mapping) for Singapore.</p><p dir="ltr">Contents: (1) 1 m shadow-model rasters on the SVY21 city grid: DEM (resampled from the ALOS PALSAR 12.5 m DEM), tree-canopy height (Meta 1 m canopy height map), building DSM, arcade box top / base, covered-linkway roof and building-footprint mask; together with the five UMEP-format meteorological forcing files of station S50 (Clementi Road) in 1_shadow_model/forcing they are the complete input of the SOLWEIG-GPU LDSM/ADSM shadow model.

(2) The 1,111 image tiles (808 training, 303 validation; 1024 x 1024 px at 0.3 m, cut from Google Earth imagery) used to train the GeoSAM + TopoLoRA covered-linkway model; their annotations and tile grid are in the code repository. (3) A pointer to the street-level imagery interface used for arcade detection (images not redistributed). (4) Transit and building-occupancy inputs of the flow models: link to the LTA DataMall passenger-volume datasets, bus stop and MRT/LRT station and exit locations, and the 20 EnergyPlus building-archetype models that provide hourly occupancy schedules.

Read the rest (8 more)

(5) Base data: building footprints with height, storeys, function and gross floor area, island boundary, and the cleaned OpenStreetMap pedestrian lines.</p><p><br></p><ul><li><b>Building footprints with height, storeys and function</b>: 118,782 footprints of Singapore with building height, storeys, archetype (function class) and gross floor area; island boundary (14 polygons). Source: City Syntax Lab building dataset (OpenStreetMap footprints; heights and functions compiled by the lab).

CC BY 4.0; footprints (c) OpenStreetMap contributors (ODbL). Used by module 1 (building DSM via module 3b), 2 (exclusion mask), 3a / 3b (arcade projection), 4a / 4b (destination weights).</li><li><b>Tree-canopy height</b>: Meta 1 m global canopy height map, clipped to Singapore (canopy height above ground). Source: Meta / WRI global canopy height maps (Tolan et al., 2024), CC BY 4.0.

Used by module 1 (CDSM, vegetation shadow).</li><li><b>Terrain</b>: ALOS PALSAR radiometrically terrain-corrected DEM (12.5 m), resampled to the 1 m city grid. Source: JAXA / METI ALOS PALSAR, distributed by ASF DAAC (free with attribution). Used by module 1 (DEM).</li><li><b>Satellite imagery, 0.3 m</b>: Google Earth imagery mosaicked to SVY21; 1,111 annotated 1024 px training tiles (808 train / 303 validation).

Source: Google (terms of use; tiles provided for research reproducibility only). Used by module 2 (covered-linkway extraction).</li><li><b>Street-level imagery, four views per point</b>: Four perspective views per panorama point every 20 m along the road network (about 148,850 points in Singapore). Source: Google Street View Static API (Google terms of use; images not redistributed).

Used by module 3a (arcade detection), 3b (projection onto building footprints).</li><li><b>Pedestrian network</b>: OpenStreetMap highway extract of Singapore (June 2026), cleaned to 404,613 pedestrian-passable segments. Source: OpenStreetMap contributors, ODbL. Used by module 5 (network reconstruction), 4b (flow model), 2 (network filter).</li><li><b>Public transport: stops, stations and passenger volumes</b>: Bus stop locations (Aug 2025), MRT / LRT stations (Aug 2025) and station exits (Feb 2025); monthly passenger volumes by bus stop and train station and by origin-destination (hourly tap-in / tap-out).

Source: Land Transport Authority of Singapore, LTA DataMall, Singapore Open Data Licence. Used by module 4b (demand origins; 4a uses its station-ridership export).</li><li><b>Building occupancy schedules</b>: 20 EnergyPlus archetype models (SGP 2025 V5): people per floor area and hourly occupancy schedules per building type. Source: Singapore building-archetype models (SGP 2025 V5).

Used by module 4b (hourly destination weights of the gravity model; 4a uses its building-weight export).</li><li><b>Meteorological forcing</b>: Hourly UMEP-format forcing of station S50 (Clementi Road): 2026-03-01 (paper run) and the four equinox / solstice days. Source: Meteorological Service Singapore; included in this record (1_shadow_model/forcing). Used by module 1 (shadow model).</li></ul><p dir="ltr">Rasters are LZW-compressed GeoTIFFs in EPSG:3414 (the canopy-height layer keeps its own extent); vectors are shapefiles or GeoPackages in EPSG:3414 (building footprints in EPSG:4326).

README.md describes every file and MANIFEST.csv gives sizes and MD5 checksums. Derived products (hourly shadow rasters, covered-linkway polygons, arcade vectors, reconstructed network and pedestrian flows) are regenerated from these inputs with the code at github.com/Shawnzhang7829/ShadeWalk (archived on Zenodo, doi:10.5281/zenodo.22688709) and are available from the corresponding author on request.</p><p dir="ltr">Licence: CC BY 4.0 for the layers compiled by the authors; OpenStreetMap-derived layers (c) OpenStreetMap contributors (ODbL); LTA layers under the Singapore Open Data Licence; the image tiles are cut from Google Earth imagery and remain subject to Google's terms of use (provided for research reproducibility only).</p>

Links

Where it is published

Catalogue records · 1

Topics

Inferred from text
Image 75% · Satellite remote sensing 65%
Provenance · 1 source records, 19 field assertions
SourceKeyLast seenRaw
figshareoai:figshare.com:article/335490255 d agoJSON v1
FieldAssertionExtractorEvidence
access_levelsource · figshare comconnector:figshare_com@1.0.0
concepts[field].anzsrc:field:330413mapping · figshare comvocabulary-mapper@1.0.0keywords['Urban planning and health']
concepts[field].local:field:astronomymapping · figshare comconnector:figshare_com@1.0.0
concepts[field].local:field:chemistrymapping · figshare comconnector:figshare_com@1.0.0
concepts[field].local:field:computer-science-aimapping · figshare comconnector:figshare_com@1.0.0
concepts[field].local:field:earth-environmentalmapping · figshare comconnector:figshare_com@1.0.0
concepts[field].local:field:economics-financemapping · figshare comconnector:figshare_com@1.0.0
concepts[field].local:field:engineeringmapping · figshare comconnector:figshare_com@1.0.0
concepts[field].local:field:humanitiesmapping · figshare comconnector:figshare_com@1.0.0
concepts[field].local:field:life-sciencesmapping · figshare comconnector:figshare_com@1.0.0
concepts[field].local:field:medicine-healthmapping · figshare comconnector:figshare_com@1.0.0
concepts[field].local:field:ocean-atmosphericmapping · figshare comconnector:figshare_com@1.0.0
concepts[field].local:field:social-sciencemapping · figshare comconnector:figshare_com@1.0.0
concepts[modality].local:modality:imageenrichment · figshare comkeyword-concept-rules@1.0.0title+description (75%)
concepts[modality].local:modality:remote-sensingenrichment · figshare comkeyword-concept-rules@1.0.0title+description (65%)
descriptionsource · figshare comconnector:figshare_com@1.0.0/metadata/dc/description
licensesource · figshare comconnector:figshare_com@1.0.0/metadata/dc/rights
publication_datesource · figshare comconnector:figshare_com@1.0.0
titlesource · figshare comconnector:figshare_com@1.0.0/metadata/dc/title